Datasets:

connections-dev
/

connection_queries_jan12_natural_original_1_reason_medium_0.7_31000_gpt-oss-120b

triple1 listlengths 3 3	triple2 listlengths 3 3	triple3 listlengths 3 3	triple1_labels listlengths 3 3	triple2_labels listlengths 3 3	triple3_labels listlengths 3 3	description stringlengths 8 19	keep_triple bool 1 class	query_prompt stringlengths 3.23k 3.36k	query stringlengths 85 226	entity_a stringlengths 3 55	entity_b stringlengths 4 85	rel_b stringclasses 88 values	index int64 0 930	prompt stringlengths 3.67k 3.89k	path_prediction listlengths 1 1	prompt_for_path listlengths 1 1	usage listlengths 1 1
[ "Q1352656", "P166", "Q935843" ]	[ "Q356715", "P166", "Q935843" ]	[ "Q356715", "P358", "Q100871850" ]	[ "Morton Gould", "award received", "Grammy Lifetime Achievement Award" ]	[ "Benny Carter", "award received", "Grammy Lifetime Achievement Award" ]	[ "Benny Carter", "discography", "Benny Carter discography" ]	Q935843__P166	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting Morton Gould, the American composer and conductor, and someone who has a notable discography in the music industry?	Morton Gould	Benny Carter discography	discography	0	Query: What are different ways of connecting Morton Gould, the American composer and conductor, and someone who has a notable discography in the music industry? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relatio...	[ "<think>We need to generate many paths from Morton Gould to Benny Carter discography, ending with a triple whose relationship is 'discography' and tail 'Benny Carter discography'. So final triple must be (some entity, discography, Benny Carter discography). That entity could be Benny Carter himself (Benny Carter (m...	[ "Query: What are different ways of connecting Morton Gould, the American composer and conductor, and someone who has a notable discography in the music industry?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, ...	[ { "completion_tokens": 3109, "prompt_tokens": 838, "total_tokens": 3947 } ]
[ "Q57067", "P166", "Q38104" ]	[ "Q131729", "P166", "Q38104" ]	[ "Q131729", "P106", "Q593644" ]	[ "Max von Laue", "award received", "Nobel Prize in Physics" ]	[ "William Lawrence Bragg", "award received", "Nobel Prize in Physics" ]	[ "William Lawrence Bragg", "occupation", "chemist" ]	Q38104__P166	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting Max von Laue, the physicist known for his work in X-ray diffraction, and someone who is recognized for their contributions as a chemist?	Max von Laue	chemist	occupation	1	Query: What are different ways of connecting Max von Laue, the physicist known for his work in X-ray diffraction, and someone who is recognized for their contributions as a chemist? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual t...	[ "<think>We need to output JSON with integer keys mapping to list of triples. Each triple: (head entity, relationship, tail entity). Start with Max von Laue, end with a triple whose relationship is 'occupation' and tail is 'chemist'. So final triple must be something like (X, occupation, chemist). X can be any chemi...	[ "Query: What are different ways of connecting Max von Laue, the physicist known for his work in X-ray diffraction, and someone who is recognized for their contributions as a chemist?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of fa...	[ { "completion_tokens": 4349, "prompt_tokens": 843, "total_tokens": 5192 } ]
[ "Q2085", "P161", "Q164328" ]	[ "Q2085", "P161", "Q174263" ]	[ "Q174263", "P19", "Q268873" ]	[ "Twin Peaks", "cast member", "David Koechner" ]	[ "Twin Peaks", "cast member", "Rodney Rowland" ]	[ "Rodney Rowland", "place of birth", "Newport Beach" ]	Q2085__P161	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting David Koechner, the American actor and comedian, and someone who was born in Newport Beach?	David Koechner	Newport Beach	place of birth	2	Query: What are different ways of connecting David Koechner, the American actor and comedian, and someone who was born in Newport Beach? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship, tail) forming a c...	[ "<think>We need to generate many distinct paths from David Koecher (actor) to someone born in Newport Beach. End condition: triple whose relationship is 'place of birth' and tail 'Newport Beach'. So path ends with (some person, place of birth, Newport Beach). That person must be \"someone who was born in Newport Be...	[ "Query: What are different ways of connecting David Koechner, the American actor and comedian, and someone who was born in Newport Beach?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship, tail) form...	[ { "completion_tokens": 5278, "prompt_tokens": 834, "total_tokens": 6112 } ]
[ "Q18037479", "P2293", "Q11081" ]	[ "Q18053558", "P2293", "Q11081" ]	[ "Q18053558", "P2548", "Q22809711" ]	[ "CD2AP", "genetic association", "Alzheimer's disease" ]	[ "ATXN7L1", "genetic association", "Alzheimer's disease" ]	[ "ATXN7L1", "strand orientation", "reverse strand" ]	Q11081__P2293	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting CD2AP, a protein involved in cellular processes, and a substance that has a reverse strand orientation?	CD2AP	reverse strand	strand orientation	3	Query: What are different ways of connecting CD2AP, a protein involved in cellular processes, and a substance that has a reverse strand orientation? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship, tail)...	[ "<think>We need to produce JSON with integer keys mapping to list of triples. Each path starts with CD2AP (a protein). Ends with a triple whose relationship is 'strand orientation' and tail is 'reverse strand'. So the final triple must be something like (X, strand orientation, reverse strand). The head X can be any...	[ "Query: What are different ways of connecting CD2AP, a protein involved in cellular processes, and a substance that has a reverse strand orientation?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship...	[ { "completion_tokens": 4659, "prompt_tokens": 834, "total_tokens": 5493 } ]
[ "Q283572", "P463", "Q254138" ]	[ "Q316136", "P463", "Q254138" ]	[ "Q316136", "P19", "Q39709" ]	[ "Kun Nong", "member of", "Slipknot" ]	[ "Chris Fehn", "member of", "Slipknot" ]	[ "Chris Fehn", "place of birth", "Des Moines" ]	Q254138__P463	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting Kun Nong and someone who was born in Des Moines?	Kun Nong	Des Moines	place of birth	4	Query: What are different ways of connecting Kun Nong and someone who was born in Des Moines? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship, tail) forming a continuous chain that begins with one entity...	[ "<think>We need to produce JSON with integer keys mapping to list of triples (head, relationship, tail). Paths start with Kun Nong and end with a triple whose relationship is 'place of birth' and tail is 'Des Moines'. So final triple must be something like (X, place of birth, Des Moines). The X is someone born in D...	[ "Query: What are different ways of connecting Kun Nong and someone who was born in Des Moines?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, relationship, tail) forming a continuous chain that begins with one...	[ { "completion_tokens": 7639, "prompt_tokens": 823, "total_tokens": 8462 } ]
[ "Q423327", "P2175", "Q12199" ]	[ "Q422654", "P2175", "Q12199" ]	[ "Q422654", "P769", "Q898407" ]	[ "enfuvirtide", "medical condition treated", "HIV/AIDS" ]	[ "saquinavir", "medical condition treated", "HIV/AIDS" ]	[ "saquinavir", "significant drug interaction", "venlafaxine" ]	Q12199__P2175	false	Task: Generate a Natural Language Query. You will be given input in the following format: Concept A: entity1 Concept B: entity2 Concept C: (concept B, rel, entity3) Concept A and Concept C are connected via an independent concept B. Concept C is a property of Concept B, which the person is supposed to guess. You...	What are different ways of connecting enfuvirtide, a medication used to treat HIV, and a substance that has a significant drug interaction with venlafaxine?	enfuvirtide	venlafaxine	significant drug interaction	5	Query: What are different ways of connecting enfuvirtide, a medication used to treat HIV, and a substance that has a significant drug interaction with venlafaxine? Task: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (head, rela...	[ "<think>We need to produce many paths connecting enfuvirtide to a substance that has a significant drug interaction with venlafaxine. So we need to find a substance that has a significant drug interaction with venlafaxine. We need a path from enfuvirtide to that substance, ending with triple (substance, significant...	[ "Query: What are different ways of connecting enfuvirtide, a medication used to treat HIV, and a substance that has a significant drug interaction with venlafaxine?\n\nTask: Identify how two real-world entities are connected by producing MANY connection paths. A connection path is a sequence of factual triples (hea...	[ { "completion_tokens": 4090, "prompt_tokens": 846, "total_tokens": 4936 } ]
[ "Q77462", "P463", "Q32645" ]	[ "Q95556", "P463", "Q32645" ]	[ "Q95556", "P451", "Q551050" ]	[ "Bela B", "member of", "Die Ärzte" ]	[ "Hagen Liebing", "member of", "Die Ärzte" ]	[ "Hagen Liebing", "unmarried partner", "Anja Caspary" ]	Q32645__P463	false	"Task: Generate a Natural Language Query.\n\nYou will be given input in the following format: \nC(...TRUNCATED)	"What are different ways of connecting Bela B, the German musician, and someone who is the unmarried(...TRUNCATED)	Bela B	Anja Caspary	unmarried partner	6	"Query: What are different ways of connecting Bela B, the German musician, and someone who is the un(...TRUNCATED)	["<think>We need to produce JSON with integer keys mapping to list of triples (head, relationship, t(...TRUNCATED)	["Query: What are different ways of connecting Bela B, the German musician, and someone who is the u(...TRUNCATED)	[ { "completion_tokens": 2885, "prompt_tokens": 837, "total_tokens": 3722 } ]
[ "Q651253", "P40", "Q1268" ]	[ "Q9356382", "P40", "Q1268" ]	[ "Q9356382", "P22", "Q11716498" ]	[ "Nicolas Chopin", "child", "Frédéric Chopin" ]	[ "Tekla Justyna Chopin", "child", "Frédéric Chopin" ]	[ "Tekla Justyna Chopin", "father", "Jakub Krzyżanowski" ]	Q1268__P40	false	"Task: Generate a Natural Language Query.\n\nYou will be given input in the following format: \nC(...TRUNCATED)	"What are different ways of connecting Nicolas Chopin, and someone who is the father of Jakub Krzyż(...TRUNCATED)	Nicolas Chopin	Jakub Krzyżanowski	father	7	"Query: What are different ways of connecting Nicolas Chopin, and someone who is the father of Jakub(...TRUNCATED)	["<think>We need to produce JSON with integer keys mapping to list of triples. Start with Nicolas Ch(...TRUNCATED)	["Query: What are different ways of connecting Nicolas Chopin, and someone who is the father of Jaku(...TRUNCATED)	[ { "completion_tokens": 2575, "prompt_tokens": 838, "total_tokens": 3413 } ]
[ "Q23505", "P40", "Q207" ]	[ "Q190628", "P40", "Q207" ]	[ "Q190628", "P39", "Q7166256" ]	[ "George H. W. Bush", "child", "George W. Bush" ]	[ "Barbara Bush", "child", "George W. Bush" ]	[ "Barbara Bush", "position held", "First Lady of the United States" ]	Q207__P40	false	"Task: Generate a Natural Language Query.\n\nYou will be given input in the following format: \nC(...TRUNCATED)	"What are different ways of connecting George H. W. Bush, the 41st President of the United States, a(...TRUNCATED)	George H. W. Bush	First Lady of the United States	position held	8	"Query: What are different ways of connecting George H. W. Bush, the 41st President of the United St(...TRUNCATED)	["<think>We need to produce JSON with integer keys mapping to list of triples (head, relationship, t(...TRUNCATED)	["Query: What are different ways of connecting George H. W. Bush, the 41st President of the United S(...TRUNCATED)	[ { "completion_tokens": 4913, "prompt_tokens": 854, "total_tokens": 5767 } ]
[ "Q18037778", "P2293", "Q131755" ]	[ "Q18046338", "P2293", "Q131755" ]	[ "Q18046338", "P5572", "Q64386" ]	[ "INTS7", "genetic association", "bipolar disorder" ]	[ "MCTP1", "genetic association", "bipolar disorder" ]	[ "MCTP1", "expressed in", "gallbladder" ]	Q131755__P2293	false	"Task: Generate a Natural Language Query.\n\nYou will be given input in the following format: \nC(...TRUNCATED)	"What are different ways of connecting INTS7, a protein involved in various cellular processes, and (...TRUNCATED)	INTS7	gallbladder	expressed in	9	"Query: What are different ways of connecting INTS7, a protein involved in various cellular processe(...TRUNCATED)	["<think>We need to produce many paths connecting INTS7 (a protein) to a substance expressed in the (...TRUNCATED)	["Query: What are different ways of connecting INTS7, a protein involved in various cellular process(...TRUNCATED)	[ { "completion_tokens": 4256, "prompt_tokens": 840, "total_tokens": 5096 } ]

End of preview. Expand in Data Studio

Dataset: connections-dev/connection_queries_jan12

This dataset was generated using the inference script with the following configuration:

Inference Parameters

Model Configuration

Model Name: openai/gpt-oss-120b
Server URL: http://localhost:9000
API Key: Not provided
Request Timeout: 300 seconds

Query Configuration

Query Type: natural
Query Column: query
Sampling Type: original

Generation Parameters

Temperature: 0.7
Max Tokens: 31000
Number of Runs: 1
Chunk Size: Entire dataset
Effort Level: medium

Feature Flags

Step by Step Reasoning: False
Reasoning Enabled: True
No Curator: False
Index Filter: None (all samples)
Start Index: None
End Index: None

Data Information

Input File: connections-dev/connection_queries_jan12
Number of Samples: 931
Output Filename: connection_queries_jan12__natural__original__1__reasoning__medium__0.7__31000__gpt-oss-120b.jsonl

Dataset Structure

The dataset contains the following key columns:

query: The original query
entity_a: Starting entity
entity_b: Target entity
rel_b: Target relationship
path_prediction: Generated connection paths
prompt_for_path: Prompts used for path generation

Usage

You can load this dataset using:

from datasets import load_dataset

dataset = load_dataset("connections-dev/connection_queries_jan12_natural_original_1_reason_medium_0.7_31000_gpt-oss-120b")

Generation Details

This dataset was generated on 2026-05-02 12:57:22 using the inference pipeline with the above configuration.

Downloads last month: 19