| { |
| "results": { |
| "unde-laudantium_lsat-rc_cot": { |
| "acc,none": 0.3345724907063197, |
| "acc_stderr,none": 0.028822264091264628, |
| "alias": "unde-laudantium_lsat-rc_cot" |
| }, |
| "unde-laudantium_lsat-lr_cot": { |
| "acc,none": 0.24705882352941178, |
| "acc_stderr,none": 0.01911709144086774, |
| "alias": "unde-laudantium_lsat-lr_cot" |
| }, |
| "unde-laudantium_lsat-ar_cot": { |
| "acc,none": 0.20869565217391303, |
| "acc_stderr,none": 0.026854108265439658, |
| "alias": "unde-laudantium_lsat-ar_cot" |
| }, |
| "unde-laudantium_logiqa_cot": { |
| "acc,none": 0.29233226837060705, |
| "acc_stderr,none": 0.018193366406024102, |
| "alias": "unde-laudantium_logiqa_cot" |
| }, |
| "unde-laudantium_logiqa2_cot": { |
| "acc,none": 0.3276081424936387, |
| "acc_stderr,none": 0.01184132971466995, |
| "alias": "unde-laudantium_logiqa2_cot" |
| }, |
| "temporibus-illo_lsat-rc_cot": { |
| "acc,none": 0.25650557620817843, |
| "acc_stderr,none": 0.026675948246675074, |
| "alias": "temporibus-illo_lsat-rc_cot" |
| }, |
| "temporibus-illo_lsat-lr_cot": { |
| "acc,none": 0.22156862745098038, |
| "acc_stderr,none": 0.018407949229981378, |
| "alias": "temporibus-illo_lsat-lr_cot" |
| }, |
| "temporibus-illo_lsat-ar_cot": { |
| "acc,none": 0.1956521739130435, |
| "acc_stderr,none": 0.026214799709819596, |
| "alias": "temporibus-illo_lsat-ar_cot" |
| }, |
| "temporibus-illo_logiqa_cot": { |
| "acc,none": 0.2364217252396166, |
| "acc_stderr,none": 0.016995363747767788, |
| "alias": "temporibus-illo_logiqa_cot" |
| }, |
| "temporibus-illo_logiqa2_cot": { |
| "acc,none": 0.30725190839694655, |
| "acc_stderr,none": 0.011639836259579922, |
| "alias": "temporibus-illo_logiqa2_cot" |
| }, |
| "quo-non_lsat-rc_cot": { |
| "acc,none": 0.26765799256505574, |
| "acc_stderr,none": 0.027044545314587293, |
| "alias": "quo-non_lsat-rc_cot" |
| }, |
| "quo-non_lsat-lr_cot": { |
| "acc,none": 0.22941176470588234, |
| "acc_stderr,none": 0.01863631913244453, |
| "alias": "quo-non_lsat-lr_cot" |
| }, |
| "quo-non_lsat-ar_cot": { |
| "acc,none": 0.1782608695652174, |
| "acc_stderr,none": 0.025291655246273914, |
| "alias": "quo-non_lsat-ar_cot" |
| }, |
| "quo-non_logiqa_cot": { |
| "acc,none": 0.2364217252396166, |
| "acc_stderr,none": 0.016995363747767788, |
| "alias": "quo-non_logiqa_cot" |
| }, |
| "quo-non_logiqa2_cot": { |
| "acc,none": 0.2926208651399491, |
| "acc_stderr,none": 0.01147864633663914, |
| "alias": "quo-non_logiqa2_cot" |
| }, |
| "magni-excepturi_lsat-rc_cot": { |
| "acc,none": 0.27137546468401486, |
| "acc_stderr,none": 0.027162503089239527, |
| "alias": "magni-excepturi_lsat-rc_cot" |
| }, |
| "magni-excepturi_lsat-lr_cot": { |
| "acc,none": 0.2411764705882353, |
| "acc_stderr,none": 0.018961774215004727, |
| "alias": "magni-excepturi_lsat-lr_cot" |
| }, |
| "magni-excepturi_lsat-ar_cot": { |
| "acc,none": 0.21304347826086956, |
| "acc_stderr,none": 0.027057754389936205, |
| "alias": "magni-excepturi_lsat-ar_cot" |
| }, |
| "magni-excepturi_logiqa_cot": { |
| "acc,none": 0.2715654952076677, |
| "acc_stderr,none": 0.017790679673144884, |
| "alias": "magni-excepturi_logiqa_cot" |
| }, |
| "magni-excepturi_logiqa2_cot": { |
| "acc,none": 0.2856234096692112, |
| "acc_stderr,none": 0.011396524130843133, |
| "alias": "magni-excepturi_logiqa2_cot" |
| }, |
| "laboriosam-numquam_lsat-rc_cot": { |
| "acc,none": 0.26765799256505574, |
| "acc_stderr,none": 0.027044545314587293, |
| "alias": "laboriosam-numquam_lsat-rc_cot" |
| }, |
| "laboriosam-numquam_lsat-lr_cot": { |
| "acc,none": 0.21764705882352942, |
| "acc_stderr,none": 0.018290217500245277, |
| "alias": "laboriosam-numquam_lsat-lr_cot" |
| }, |
| "laboriosam-numquam_lsat-ar_cot": { |
| "acc,none": 0.1782608695652174, |
| "acc_stderr,none": 0.025291655246273914, |
| "alias": "laboriosam-numquam_lsat-ar_cot" |
| }, |
| "laboriosam-numquam_logiqa_cot": { |
| "acc,none": 0.23482428115015974, |
| "acc_stderr,none": 0.016955557820725036, |
| "alias": "laboriosam-numquam_logiqa_cot" |
| }, |
| "laboriosam-numquam_logiqa2_cot": { |
| "acc,none": 0.2926208651399491, |
| "acc_stderr,none": 0.011478646336639108, |
| "alias": "laboriosam-numquam_logiqa2_cot" |
| }, |
| "dolore-possimus_lsat-rc_cot": { |
| "acc,none": 0.31226765799256506, |
| "acc_stderr,none": 0.028307781204694345, |
| "alias": "dolore-possimus_lsat-rc_cot" |
| }, |
| "dolore-possimus_lsat-lr_cot": { |
| "acc,none": 0.2529411764705882, |
| "acc_stderr,none": 0.019267629016819672, |
| "alias": "dolore-possimus_lsat-lr_cot" |
| }, |
| "dolore-possimus_lsat-ar_cot": { |
| "acc,none": 0.2826086956521739, |
| "acc_stderr,none": 0.02975452853823325, |
| "alias": "dolore-possimus_lsat-ar_cot" |
| }, |
| "dolore-possimus_logiqa_cot": { |
| "acc,none": 0.2939297124600639, |
| "acc_stderr,none": 0.01822240539964835, |
| "alias": "dolore-possimus_logiqa_cot" |
| }, |
| "dolore-possimus_logiqa2_cot": { |
| "acc,none": 0.3428753180661578, |
| "acc_stderr,none": 0.011975782754482172, |
| "alias": "dolore-possimus_logiqa2_cot" |
| } |
| }, |
| "configs": { |
| "dolore-possimus_logiqa2_cot": { |
| "task": "dolore-possimus_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "dolore-possimus-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "dolore-possimus_logiqa_cot": { |
| "task": "dolore-possimus_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "dolore-possimus-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "dolore-possimus_lsat-ar_cot": { |
| "task": "dolore-possimus_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "dolore-possimus-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "dolore-possimus_lsat-lr_cot": { |
| "task": "dolore-possimus_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "dolore-possimus-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "dolore-possimus_lsat-rc_cot": { |
| "task": "dolore-possimus_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "dolore-possimus-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "laboriosam-numquam_logiqa2_cot": { |
| "task": "laboriosam-numquam_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "laboriosam-numquam-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "laboriosam-numquam_logiqa_cot": { |
| "task": "laboriosam-numquam_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "laboriosam-numquam-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "laboriosam-numquam_lsat-ar_cot": { |
| "task": "laboriosam-numquam_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "laboriosam-numquam-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "laboriosam-numquam_lsat-lr_cot": { |
| "task": "laboriosam-numquam_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "laboriosam-numquam-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "laboriosam-numquam_lsat-rc_cot": { |
| "task": "laboriosam-numquam_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "laboriosam-numquam-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "magni-excepturi_logiqa2_cot": { |
| "task": "magni-excepturi_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "magni-excepturi-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "magni-excepturi_logiqa_cot": { |
| "task": "magni-excepturi_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "magni-excepturi-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "magni-excepturi_lsat-ar_cot": { |
| "task": "magni-excepturi_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "magni-excepturi-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "magni-excepturi_lsat-lr_cot": { |
| "task": "magni-excepturi_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "magni-excepturi-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "magni-excepturi_lsat-rc_cot": { |
| "task": "magni-excepturi_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "magni-excepturi-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "quo-non_logiqa2_cot": { |
| "task": "quo-non_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "quo-non-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "quo-non_logiqa_cot": { |
| "task": "quo-non_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "quo-non-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "quo-non_lsat-ar_cot": { |
| "task": "quo-non_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "quo-non-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "quo-non_lsat-lr_cot": { |
| "task": "quo-non_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "quo-non-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "quo-non_lsat-rc_cot": { |
| "task": "quo-non_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "quo-non-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "temporibus-illo_logiqa2_cot": { |
| "task": "temporibus-illo_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "temporibus-illo-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "temporibus-illo_logiqa_cot": { |
| "task": "temporibus-illo_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "temporibus-illo-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "temporibus-illo_lsat-ar_cot": { |
| "task": "temporibus-illo_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "temporibus-illo-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "temporibus-illo_lsat-lr_cot": { |
| "task": "temporibus-illo_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "temporibus-illo-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "temporibus-illo_lsat-rc_cot": { |
| "task": "temporibus-illo_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "temporibus-illo-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "unde-laudantium_logiqa2_cot": { |
| "task": "unde-laudantium_logiqa2_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "unde-laudantium-logiqa2/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "unde-laudantium_logiqa_cot": { |
| "task": "unde-laudantium_logiqa_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "unde-laudantium-logiqa/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "unde-laudantium_lsat-ar_cot": { |
| "task": "unde-laudantium_lsat-ar_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "unde-laudantium-lsat-ar/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "unde-laudantium_lsat-lr_cot": { |
| "task": "unde-laudantium_lsat-lr_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "unde-laudantium-lsat-lr/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| }, |
| "unde-laudantium_lsat-rc_cot": { |
| "task": "unde-laudantium_lsat-rc_cot", |
| "group": "logikon-bench", |
| "dataset_path": "logikon/cot-eval-traces", |
| "dataset_kwargs": { |
| "data_files": { |
| "test": "unde-laudantium-lsat-rc/test-00000-of-00001.parquet" |
| } |
| }, |
| "test_split": "test", |
| "doc_to_text": "def doc_to_text_cot(doc) -> str:\n \"\"\"\n Answer the following question about the given passage. [Base your answer on the reasoning below.]\n \n Passage: <passage>\n \n Question: <question>\n A. <choice1>\n B. <choice2>\n C. <choice3>\n D. <choice4>\n [E. <choice5>]\n \n [Reasoning: <reasoning>]\n \n Answer:\n \"\"\"\n k = len(doc[\"options\"])\n choices = [\"a\", \"b\", \"c\", \"d\", \"e\"][:k]\n prompt = \"Answer the following question about the given passage. Base your answer on the reasoning below.\\n\\n\"\n prompt = \"Passage: \" + doc[\"passage\"] + \"\\n\\n\"\n prompt += \"Question: \" + doc[\"question\"] + \"\\n\"\n for choice, option in zip(choices, doc[\"options\"]):\n prompt += f\"{choice.upper()}. {option}\\n\"\n prompt += \"\\n\"\n prompt += \"Reasoning: \" + doc[\"reasoning_trace\"] + \"\\n\\n\" \n prompt += \"Answer:\"\n return prompt\n", |
| "doc_to_target": "{{answer}}", |
| "doc_to_choice": "{{options}}", |
| "description": "", |
| "target_delimiter": " ", |
| "fewshot_delimiter": "\n\n", |
| "num_fewshot": 0, |
| "metric_list": [ |
| { |
| "metric": "acc", |
| "aggregation": "mean", |
| "higher_is_better": true |
| } |
| ], |
| "output_type": "multiple_choice", |
| "repeats": 1, |
| "should_decontaminate": false, |
| "metadata": { |
| "version": 0.0 |
| } |
| } |
| }, |
| "versions": { |
| "dolore-possimus_logiqa2_cot": 0.0, |
| "dolore-possimus_logiqa_cot": 0.0, |
| "dolore-possimus_lsat-ar_cot": 0.0, |
| "dolore-possimus_lsat-lr_cot": 0.0, |
| "dolore-possimus_lsat-rc_cot": 0.0, |
| "laboriosam-numquam_logiqa2_cot": 0.0, |
| "laboriosam-numquam_logiqa_cot": 0.0, |
| "laboriosam-numquam_lsat-ar_cot": 0.0, |
| "laboriosam-numquam_lsat-lr_cot": 0.0, |
| "laboriosam-numquam_lsat-rc_cot": 0.0, |
| "magni-excepturi_logiqa2_cot": 0.0, |
| "magni-excepturi_logiqa_cot": 0.0, |
| "magni-excepturi_lsat-ar_cot": 0.0, |
| "magni-excepturi_lsat-lr_cot": 0.0, |
| "magni-excepturi_lsat-rc_cot": 0.0, |
| "quo-non_logiqa2_cot": 0.0, |
| "quo-non_logiqa_cot": 0.0, |
| "quo-non_lsat-ar_cot": 0.0, |
| "quo-non_lsat-lr_cot": 0.0, |
| "quo-non_lsat-rc_cot": 0.0, |
| "temporibus-illo_logiqa2_cot": 0.0, |
| "temporibus-illo_logiqa_cot": 0.0, |
| "temporibus-illo_lsat-ar_cot": 0.0, |
| "temporibus-illo_lsat-lr_cot": 0.0, |
| "temporibus-illo_lsat-rc_cot": 0.0, |
| "unde-laudantium_logiqa2_cot": 0.0, |
| "unde-laudantium_logiqa_cot": 0.0, |
| "unde-laudantium_lsat-ar_cot": 0.0, |
| "unde-laudantium_lsat-lr_cot": 0.0, |
| "unde-laudantium_lsat-rc_cot": 0.0 |
| }, |
| "n-shot": { |
| "dolore-possimus_logiqa2_cot": 0, |
| "dolore-possimus_logiqa_cot": 0, |
| "dolore-possimus_lsat-ar_cot": 0, |
| "dolore-possimus_lsat-lr_cot": 0, |
| "dolore-possimus_lsat-rc_cot": 0, |
| "laboriosam-numquam_logiqa2_cot": 0, |
| "laboriosam-numquam_logiqa_cot": 0, |
| "laboriosam-numquam_lsat-ar_cot": 0, |
| "laboriosam-numquam_lsat-lr_cot": 0, |
| "laboriosam-numquam_lsat-rc_cot": 0, |
| "magni-excepturi_logiqa2_cot": 0, |
| "magni-excepturi_logiqa_cot": 0, |
| "magni-excepturi_lsat-ar_cot": 0, |
| "magni-excepturi_lsat-lr_cot": 0, |
| "magni-excepturi_lsat-rc_cot": 0, |
| "quo-non_logiqa2_cot": 0, |
| "quo-non_logiqa_cot": 0, |
| "quo-non_lsat-ar_cot": 0, |
| "quo-non_lsat-lr_cot": 0, |
| "quo-non_lsat-rc_cot": 0, |
| "temporibus-illo_logiqa2_cot": 0, |
| "temporibus-illo_logiqa_cot": 0, |
| "temporibus-illo_lsat-ar_cot": 0, |
| "temporibus-illo_lsat-lr_cot": 0, |
| "temporibus-illo_lsat-rc_cot": 0, |
| "unde-laudantium_logiqa2_cot": 0, |
| "unde-laudantium_logiqa_cot": 0, |
| "unde-laudantium_lsat-ar_cot": 0, |
| "unde-laudantium_lsat-lr_cot": 0, |
| "unde-laudantium_lsat-rc_cot": 0 |
| }, |
| "config": { |
| "model": "vllm", |
| "model_args": "pretrained=mistralai/Mistral-7B-v0.1,revision=main,dtype=auto,tensor_parallel_size=1,gpu_memory_utilization=0.9,trust_remote_code=true,max_length=4096", |
| "batch_size": "auto", |
| "batch_sizes": [], |
| "device": null, |
| "use_cache": null, |
| "limit": null, |
| "bootstrap_iters": 100000, |
| "gen_kwargs": null |
| }, |
| "git_hash": "5044cf9" |
| } |
