Polish dataset card and add centered project artwork

README.md

@@ -86,7 +86,13 @@ configs:
   </div>
 </div>
 
-![Bactrainus modular architecture](architecture.svg)
+<p align="center">
+  <img src="assets/data.png" alt="Bactrainus HotpotQA training-data artwork" width="760">
+</p>
+
+<p align="center">
+  <img src="architecture.svg" alt="Bactrainus modular architecture" width="960">
+</p>
 
 ## At a glance
 
@@ -112,6 +118,7 @@ dataset = load_dataset(
     "bactrianus/bactrainus-hotpotqa",
     "cot-reader-sft",
     split="train",
+    revision="v1.0.0",
 )
 
 print(dataset.num_rows)          # 90447
@@ -119,14 +126,14 @@ print(dataset[0]["source_id"])
 print(dataset[0]["messages"])
 ```
 
-For a reproducible experiment, pin the dataset commit:
+For a reproducible experiment, pin the immutable release tag:
 
 ```python
 dataset = load_dataset(
     "bactrianus/bactrainus-hotpotqa",
     "structured",
     split="train",
-    revision="7f3a1d4d21f22aad7262d8ffd6520f31186b284d",
+    revision="v1.0.0",
 )
 ```
 
@@ -268,10 +275,16 @@ The IDs make it straightforward to combine configurations without fuzzy joins:
 
 ```python
 reader = load_dataset(
-    "bactrianus/bactrainus-hotpotqa", "reader-sft", split="train"
+    "bactrianus/bactrainus-hotpotqa",
+    "reader-sft",
+    split="train",
+    revision="v1.0.0",
 )
 decomposer = load_dataset(
-    "bactrianus/bactrainus-hotpotqa", "question-decomposer-sft", split="train"
+    "bactrianus/bactrainus-hotpotqa",
+    "question-decomposer-sft",
+    split="train",
+    revision="v1.0.0",
 )
 
 assert set(reader["source_id"]) == set(decomposer["source_id"])
@@ -304,7 +317,9 @@ attribution and use a compatible ShareAlike license. See
 
 ## Citation
 
-If you use this training suite, cite both Bactrainus and HotpotQA:
+If you use this training suite, cite the Bactrainus article. Retain the
+upstream HotpotQA attribution described in [`ATTRIBUTION.md`](ATTRIBUTION.md)
+when redistributing or adapting the data.
 
 ```bibtex
 @article{barati2025bactrainus,
@@ -314,15 +329,6 @@ If you use this training suite, cite both Bactrainus and HotpotQA:
   year    = {2025},
   url     = {https://arxiv.org/abs/2501.06286}
 }
-
-@inproceedings{yang2018hotpotqa,
-  title     = {{HotpotQA}: A Dataset for Diverse, Explainable Multi-hop Question Answering},
-  author    = {Yang, Zhilin and Qi, Peng and Zhang, Saizheng and Bengio, Yoshua and Cohen, William W. and Salakhutdinov, Ruslan and Manning, Christopher D.},
-  booktitle = {Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing},
-  year      = {2018},
-  pages     = {2369--2380},
-  doi       = {10.18653/v1/D18-1259}
-}
 ```
 
 ## Links