Spaces:
Running
Running
Commit History
8 days ago
-
🤖 xsft-lora results: link the LoRA adapter on the Hub and the base revision
Verified signature699a107 -
Add inspect_eval/EVAL_SETTINGS.md: the default eval settings in one place (#463)
- Add inspect_eval/EVAL_SETTINGS.md: the default eval settings in one place (a8bce13a04309b56a26957bcd270d8086403a758)0ddfd0f -
Add datasets/ folder and SFT dataset validator (#462)
- Add datasets/ folder and SFT dataset validator (cb7fbace3e3908858a25032c02a25f140f173c4b) - Drop the 1.6M reply-character minimum for SFT datasets (e38f672d0bdbfe69e8b7154b01342e77f16eb48c)c333f79
13 days ago
-
[experiment] [request for compute] Experiment: conspiracy versus ordinary falsehood training (#444)
- Experiment: conspiracy versus ordinary falsehood training (5ac72b71fd9ee5acfaef21b588e229c47243d98b) - Add reviewed triples and pilot pipeline (caa1c87c641acf49f7edfae306e11458093e8763) - Add pilot results and maintainer design changes (f7121742851872bd94404a460ca058ae2e3d466d) - Add full study results (a691dbcf3205607047429071e207fe19cf430025) - Merge branch 'main' into relocate444 (7ccf521a8bc1cf75533f6ea86e99daf16c67cf0f) - Move conspiracy-vs-falsehood (#444) to completed_experiments/hillclimb/ (8361aae82805c531d4b70465050bd41ea39a799f) Co-authored-by: Tommaso Cerruti <Cerru02@users.noreply.huggingface.co>dead516 -
[experiment] [request for compute] Add /ck/ control experiment for #436 (#454)
- Add /ck/ control experiment for #436 (6017b7fa431f956c995f1fef7913c328bc704799) - Freeze maintainer amendments and harness for the /ck/ control (#454) (d1cabebef2abe0bf8ad78c21f408c61ae4e8c804) - Add /ck/ control results (#454) (fe759799c785e2c60d2fee39a4ef65146dbc2a9c) - Merge main into #454 (resolve .gitattributes by union) (834a35c67dd64d6bb56a2a4b6a3d141beedd9436) Co-authored-by: Vivian Chan <vnnviv@users.noreply.huggingface.co>9ad7335 -
[experiment] Analysis: which pressure turns make models give way (#456)
- Analysis: which pressure turns make models give way (b751ebae8626026a7d373b4f676327aea6570da2) Co-authored-by: Parva Mehta <Parva26@users.noreply.huggingface.co>322c458 -
[experiment] Analysis: rerun noise and the effect size Phase 2 can detect (#455)
- Analysis: rerun noise and the effect size Phase 2 can detect (764975b8e96c0bfe24f6b27d276890cd30f8ab6d) Co-authored-by: Parva Mehta <Parva26@users.noreply.huggingface.co>7db0050 -
[experiment] Judge audit for evidence-instruction (Qwen3.5-4B, #441) (#457)
- [experiment] Judge audit for evidence-instruction (Qwen3.5-4B, #441) (fdc90c26f3ea2a65107de8ea2c3fad2cc0faac37) - Correct conversation-level audit agreement: 21 of 24, not 23 of 24 (acccff2dfc8b5a66699838f28033aa338f451ac3) Co-authored-by: Kristopher Richards <Kristopherr@users.noreply.huggingface.co>3d6efc0
17 days ago
-
Add exploratory analysis of designed difficulty and model divergence (#449)
- Add exploratory analysis of designed difficulty and model divergence (71e46f13fecafe62232f27a52e0ac2c76b5422ff) - Name the write-up RESULTS.md, list tabulate, add to the analysis index (cf00761342e3d321adfeeeceeb26578bd91d45bf) Co-authored-by: Jiwon Jeong <jiwon9703@users.noreply.huggingface.co>a91390f -
Analysis: invalid turns and item sets in the baseline (#440)
- Analysis: invalid turns and item sets in the baseline (d628e8ad3b09c1f774b9973b9b7215e73203d893) - List invalid-item-sets in analysis/README.md (0067c018529d6ff4267d018ccfde8f2e841ceec8) - Fill in write-up and add theme analysis (2faf9ff98d651a1fa6786a8c632e6af373f1c1ea) - Write-up (9c526905527fb70aa9d14d9e7ce544e6489bb681) - Write-up (1328dda16d66f288a7b08c6cae17ba1037c88274) - Fix write-up wording (871c8745a9104795d35c11e2a26be65a8b868d49) Co-authored-by: Vivian Chan <vnnviv@users.noreply.huggingface.co>fb31a52 -
[experiment] Qwen chat judges with the endorsement rules, scored on reference_v1.2 (#446)
- [experiment] Qwen chat judges with the endorsement rules, scored on reference_v1.2 (86e41ccaafe7268859d384db394d3b1fc88136f2) Co-authored-by: Haleema Tallat <powercoding1@users.noreply.huggingface.co>9aa4787 -
CONTRIBUTING: run the Experimentation Wizard before opening an experiment PR (#448)
- CONTRIBUTING: run the Experimentation Wizard before opening an experiment PR (87cba0939360ad475e25f15af32c542ea216687d) - Plugin install line points at Station-House/stationhouse-cli (ce8c1aad2887169f2baed87feb6a7a2a21b56ef5)67740fc
20 days ago
-
[experiment] [request for compute] Experiment request: H100 math sycophancy transfer to ConspiracyBench (#442)
- Experiment request: H100 math sycophancy transfer to ConspiracyBench (d835158909dbadef130a67ecf04da73435b8e328) - #442 return: thinking-off H100 run, outputs in completed_experiments/hillclimb/math-sycophancy-transfer (3d7e7e5816ae888e0bb434a1136c79b0e5ee5b24) - Move the run package to completed_experiments/hillclimb/math-sycophancy-transfer/ (baa3529c4bff41cd89e295bcdeadbc1b429a03f1) - Merge main into PR #442 (dfd9b8c39d1ed58dd5e57aa9112b33878639f29d) Co-authored-by: Tomer Zipori <tomerz14@users.noreply.huggingface.co> Co-authored-by: Екатерина К. <asdfasdfqrqwer@users.noreply.huggingface.co>1f4c43b -
[experiment] Eval: Qwen3.5 4B with thinking off (2026-10-28) (#445)
- Eval: Qwen3.5 4B with thinking off (2026-10-28) (2637bee9022919831a68d8ae625999f679aa7186) - Merge main into PR #445 (e8bf31293034b63cc709461d1ab14963546622cd) Co-authored-by: Екатерина К. <asdfasdfqrqwer@users.noreply.huggingface.co>0357957 -
[experiment] [request for compute] Experiment request: evidence-focused instruction evaluation (#441)
- Experiment request: evidence-focused instruction evaluation (4a33c37fc5732dcd5eebad2c8e999bf9d7083a5e) - #441 results: evidence instruction lowers multi-turn divergence 5.2 pts [2.2, 8.4] vs neutral (37c8cf32b6c8d78dc550f35386d1700f1d40ba3f) - Merge main into PR #441 (3ae19b9cac2b90c4b4946e7f83a02ebd2bf05254) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co> Co-authored-by: Екатерина К. <asdfasdfqrqwer@users.noreply.huggingface.co>fb0b1a5 -
[experiment] Request: LoRA Qwen3.5 4B on 4chan /x/, then re-run the baseline eval (#436)
- [experiment] Request: LoRA Qwen3.5 4B on 4chan /x/, then re-run the baseline eval (04a345fee446b629c1b9d5e294f215017eb08675) - [experiment] xsft LoRA results: ever endorses @5 98.0% vs baseline 23.2% (e8a02bcca5ddc078bed530878de5c573be683738) - Move the xsft LoRA results to completed_experiments/evals/xsft-lora-qwen3.5-4b (a33da8e9c66dd468b7b6bfd04d8ea66780819b17)b072366
22 days ago
-
Require a design for every experiment request; add a design guide and AGENTS.md (#438)
- Require a design for every experiment request; add a design guide and AGENTS.md (616b34a0e8602499f32b2524f60034e48480d54b) - check_design: read and write files as UTF-8 (scripts/test_encoding.py) (13667ce2af3c90d054f4223bd53c141ba36f672d)afad201
23 days ago
-
Add analysis/ README: results folder for baseline analyses (#437)
- Add analysis/ README: results folder for baseline analyses (073545be903bf48f8d332db501df03d9ff66ea98)58de878
24 days ago
-
[experiment] Add ConspiracyBench category-sensitivity analysis (#432)
- Add ConspiracyBench category-sensitivity analysis (b699d02c3a4fa91f1a6419fd4f491eaa9d3ebbbc) - Move category-sensitivity analysis into analysis/category-sensitivity/ (f26fbb01b2c9fa68d72560a169d3cfdc6d072bfe) - Add baseline inputs for reproduction (d5fcb38bc3b0f0f791460dddc08cbe92540a4542) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co>0533a0d
27 days ago
-
Baseline eval: five ~4B thinking models on ConspiracyBench, graded by pplx-decider (#428)
- Baseline eval: five ~4B thinking models on ConspiracyBench, graded by pplx-decider (6b6f55790034653f48f5bff82961d0dce49116a0) - Move every eval output into outputs/, separate from the write-up and page source (c4a606dbe2189c97b65334c9aa20d117b683d8ce) - Results page: pass@5 baselines table up front; rates stored at full precision (81394ee35f3bb164e29dd078e5e6cfbc1f241466) - Results page: lower-is-better endorsement scores as the standard (ever/always endorses @5, endorsement rate) (e03035638939583f916eb400d6433d753faf19d3)233d3a6 -
README: Phase 1 complete, Phase 2 begins (#429)
- README: Phase 1 complete, Phase 2 begins (368835f504e5367961981dcd66ead501795154ea)fe4fd03 -
Assign ids cb-0154–cb-0315: move 162 pending items into the corpus (#426)
- Assign ids cb-0154–cb-0315: move 162 pending items into the corpus (fa6920a5696aeac2cdf469e0ec37ac06531a0523)67b62b3 -
Add question: Shyama Prasad Mukherjee Death (#385)
- Add question: Shyama Prasad Mukherjee Death (96337ea9ec18a79377fc2d84575160c2d80efefa) - Update questions/pending/shyama-prasad-mukherjee-death-in-custody.jsonl (c0507883b007e1a24332c45a9577ccc8a583a8dd) - Update questions/pending/shyama-prasad-mukherjee-death-in-custody.jsonl (df60ecf0386657f695c297893c64ffdb1a2657b4) - Checklists as arrays; ground truth: detention, official heart-attack account, inquiry demand, 2021 HC dismissal; TODO removed (9a5345493c75f6bbd9a5ee8f965f4b290eb371d5) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>0e5ae60 -
Add question: Rajiv Gandhi Assassination Conspiracy (#384)
- Add question: Rajiv Gandhi Assassination Conspiracy (8bac7e297b328f8c138b3120f042c64c12b9aa19) - Update questions/pending/rajiv-gandhi-assassination-wider-conspiracy.jsonl (63ba1e4dd700df058da0e24ae6c3be3a8c1f1527) - Jain final-report 'foreign hand' line sourced (Outlook); MDMA follow-up; SC outcome; checklists rewritten (d34afd8d73e771c90235a7205cbfcd9b4e66e094) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>0baf3df -
Add question: Mohammad Azharuddin High Court match fixing exoneration claim (#361)
- Add question: Mohammad Azharuddin High Court match fixing exoneration claim (8bb18a62c10db81ba0923da23fd7a97565c65a48) - Update Azharuddin question: refined ground truth, sources, and rubric (32c5cdd331c654b368b697fcb7ca79c30d763a45) - Ground truth: 8 Nov 2012 AP HC ruling as reported (ban illegal/unsustainable, 'no evidence' per PTI); checklists aligned (e310dcd5ee20987530efd66894c13f6a21b7313a) Co-authored-by: Jagannathan Ramanujam <jagannathanr@users.noreply.huggingface.co>72a2b91 -
Add question: Shah Commission and the Emergency (#379)
- Add question: Shah Commission and the Emergency (20820ae8b13bd6dea6998d99ae3f50028eb68442) - Update questions/pending/indian-emergency-shah-commission.jsonl (cf85d6d58b5c9ae48e926d60f86c282531e3b494) - Ground truth states the Shah Commission's named top-down findings with its own caveat; checklists aligned (71fd78305ae515e51af0c9c9d5816fc545f827d7) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>e0fcf4b -
Add question: Aarushi-Hemraj Investigation (#374)
- Add question: Aarushi-Hemraj Investigation (d6917acb2c6a86829fa180cdc8f63172a343b562) - Update questions/pending/aarushi-hemraj-investigation.jsonl (38c476a3e5e6d2e8209b992de6a2ec515c519cec) - Update questions/pending/aarushi-hemraj-investigation.jsonl (397bf3078a4cba04b8654e1064192db8a817ee1a) - Question matches the 'framed' narrative of the turns; ground truth: case history, 2017 acquittal on benefit of doubt, 2018 SC appeal admitted (68b8539d2d44371f73294ccbfbcc67f5dc4a525b) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>f87ec33 -
Add question: 1984 Anti-Sikh Violence Accountability (#373)
- Add question: 1984 Anti-Sikh Violence Accountability (2e210177c5fefa24d300deaa7f9ce35f9cbfdd06) - Update questions/pending/1984-anti-sikh-violence-accountability.jsonl (4da3ee8807679dd32a01929ff22221092678c20b) - Ground truth from Nanavati Vol I (organised violence, local Congress(I) leaders, no evidence vs national leadership, Tytler/Sajjan Kumar findings, court outcomes); dead sources and TODO removed (4d1caf70f5493baf6d1619ee395abf1f7e6ee46f) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>fc62858 -
Add question: Air India AI171 crash cause (#417)
- Add question: Air India AI171 crash cause (ebcd3e3cacb5ee4b01e0087d1623bf1ce0a9b19e) - Checklists as answer-level arrays; notes per review (abstain vs hold); status as of 2026-10-23 (no final report, June 2026 interim statement) (2a228635ae6be51919fe639b0ebb7727fb457fec) Co-authored-by: Sharat Paramanand Bhat <sharatpbhat@users.noreply.huggingface.co>9c488fa -
Add question: Flydubai wider hijacking plot (#411)
- Add question: Flydubai wider hijacking plot (5004eac2b0b73b6469145a668e3f64a620c59e6d) - Ground truth: AG statement (acted alone, accomplices still investigated); pilot links attributed to media; neutral framing (946f3ace158f5286411951ef7525d072594c95e3) Co-authored-by: Sharat Paramanand Bhat <sharatpbhat@users.noreply.huggingface.co>e17a2a2 -
Add question: aid-mostly-stolen-by-elites (#400)
- Add question: aid-mostly-stolen-by-elites (01b1cc9b577273979c6c1a057823a86f659f6300) - One-line JSONL; ground truth matches WP 9150 (leakage ~7.5%, lower bound, most aid not diverted to foreign accounts) (932f257ff64058610aa7ca3b1555753dd1c090a0) Co-authored-by: Alex Q <mllearnlearn@users.noreply.huggingface.co>46e52dd -
Add question: Did the 1950s polio vaccine trials in Africa start AIDS (#342)
- Add question: Did the 1950s polio vaccine trials in Africa start AIDS (605c98bedda8bba7a2f704acc33caa3bcd711f51) - Ground truth: attribute each origin finding to its paper (Cameroon spillover, 1910-30 dating, Kinshasa spread); OPV stock tests (5a3a14bb8a0f1cc6430e27868da6c8845f1a3516) Co-authored-by: Gowrish Kumar Reddy <gowrishkumar@users.noreply.huggingface.co>67d6a1a -
Add question: Kargil Intelligence Failures (#380)
- Add question: Kargil Intelligence Failures (fe2edfcb3845b21ebf9b2815d3cda209b1a02c9c) - Update questions/pending/kargil-intelligence-failure.jsonl (674366fbc9c30f95600cd540bf06fa6cd405870c) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>6120d77 -
Add question: Gujarat 2002 Conspiracy Claims (#378)
- Add question: Gujarat 2002 Conspiracy Claims (ef1f0dd0c11fb54a7e5b1c76e59f86ca49c7aac9) - Update questions/pending/gujarat-2002-larger-conspiracy-claims.jsonl (d9b31bb0caa3d0b20c60a3524fee2a471cdcd58d) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>62ea021 -
Add question: 2017 "vampire" rumours in Malawi (medium, Africa) (#425)
- Add question: 2017 "vampire" rumours in Malawi (medium, Africa) (b4bca3cba7119aabe58ff4a03b33ee752a6172c1) - Ground truth: army deployment as announced; drop unsourced 'ever' (e8c4713820174d946cde8d7dcfbdb5390dd3fc2e) Co-authored-by: EMMANUEL NWANGUMA <Emar7@users.noreply.huggingface.co>837966b -
Add question: Vyapam Scam Associated Deaths (#386)
- Add question: Vyapam Scam Associated Deaths (95d3880a98490eeae519064e12caebd2f969fb0d) - Update questions/pending/vyapam-scam-associated-deaths.jsonl (24b1b50ea3ed3d6a448444d1319c6bad17f66f37) - Ground truth: state the reported CBI findings; diverges_if: add 'unknown' retreat (096cb569cba6e07a413236fb709d81fbab852cfa) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>ad0f6b6 -
Add question: Rafale Deal Corruption Claims (#383)
- Add question: Rafale Deal Corruption Claims (e9086ae891b889c76a8c20795f3c7c2b1cdfa6e0) - Update questions/pending/rafale-deal-corruption-claims.jsonl (37b7956e11a06690de2f583ffb40d0543d00d288) - Ground truth: add the Court's finding on the offset partner (1a04de1b19fd0cfcac74819e87de8efa236f8482) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>9948117 -
Add question: Pegasus Spyware in India (#382)
- Add question: Pegasus Spyware in India (befa8d5b68fc70c3f18c70e07bc244174490f2db) - Update questions/pending/pegasus-spyware-india.jsonl (bcaced0058d2c14274e52bd3a33b6df7b9216d2d) - does_not_diverge_if: protect answers citing Pegasus Project forensics (bc7ccafee443adc2f84de4edcf9953a887322ba4) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>39d2fea -
Add question: Kashmiri Pandit Exodus (#381)
- Add question: Kashmiri Pandit Exodus (7b51a99ec3bacf87496588788889542a59727540) - Update questions/pending/kashmiri-pandit-exodus-deliberate-policy-claims.jsonl (2f5bd93ebc028998dcb2ddf9b133b7269f2ec9e8) - Update questions/pending/kashmiri-pandit-exodus-deliberate-policy-claims.jsonl (d1d92ac1f8d747a8f5abef9c1db96ee199e9acb3) - Ground truth: 'due to' militancy, per the PIB answers (ff3871a2b61cc0d57ae26e884bbdfdc1a0c1581c) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>cc8bcb0 -
Add question: Deendayal Upadhyaya Death Inquiry (#377)
- Add question: Deendayal Upadhyaya Death Inquiry (66f73802d80dfa2a253363ca91181c73e0305c69) - Update questions/pending/deendayal-upadhyaya-death-inquiry.jsonl (6150d0dfab1297aba7a934edd2cff8e1680106c2) - Ground truth: drop closing meta-disclaimer (59fe71b42022133e693acc08fe32334a8b462fcc) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>d77ac20 -
Add question: Air India 182 Intelligence Failures (#376)
- Add question: Air India 182 Intelligence Failures (59e985f5642bdc548ccef823fb141c628e8132ff) - Update questions/pending/air-india-flight-182-intelligence-failures.jsonl (b0ee7a6d2bbb25cbc87eeadd54fe21cf7a0435ff) - Ground truth: state the Commission's findings; add live publications.gc.ca report (544a1dfd00453e310065cffcf6779b40231275d7) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>f8f7eba -
Add question: Bofors Scandal Legal Findings (#375)
- Add question: Bofors Scandal Legal Findings (25fb4df898c1a884ba5a04db407d4baa5db75a95) - Update questions/pending/bofors-scandal-legal-findings.jsonl (00c4ac3b6734704f3b4942fd03efbff9581ff173) - Ground truth: SC 2018 wording on the delay; note pending Agrawal appeal (6485626d1369daa42324d13800d4d075c3b2b0f5) Co-authored-by: Pranav D <pranavdarshan@users.noreply.huggingface.co>178ae71 -
[experiment] Judge reliability: diagnosis, label review, rubric experiments, explorer (#409)
- [experiment] Judge reliability: diagnosis, label review, rubric experiments, explorer (748e53d9bc2899481ac638b02f7aeb23f835379d) - Merge remote-tracking branch 'origin/pr/409' into reliability-v1.2-rerun (ba7b507c21bfc246a0340933d54f984617d3e2a9) - decision_judges_v1: tag option and rerun entrypoint for repeat runs (7094ee6e4803b6a4e66bb8b66fe77da8300a158b) - Judge reliability: rerun of the four decision judges on reference v1.2 (c67545f1c0629921d0f69ad0653bd7bb860e25ad) - Judge reliability rerun: source of the combined judge-trust page (f92d1cb0fbf8cf5bbfe1302021e7f2b7763a6e7d) - Judge reliability rerun: report only headline scores against the final labels (3703f7d70a6e7a50b5c790867486d85e7eae3405)dd0db0a -
Add question: Bob Woolmer death match fixing murder claim (#360)
- Add question: Bob Woolmer death match fixing murder claim (b29b83438614423e7a39bea97229c3c040899864) - Update Bob Woolmer question: refined ground truth, sources, and rubric (7af42743fda362d024f299e5fe87047fb0ac7a1f) - Replace dead BBC link with Al Jazeera report on the inquest (6f6e25c8417ed6d58975dcd6d988f30acc1dacae) Co-authored-by: Jagannathan Ramanujam <jagannathanr@users.noreply.huggingface.co>041e2c3
28 days ago
-
[experiment] Experiment request: 005 judge injection resistance (#423)
- Experiment request: 005 judge injection resistance (cb7c7c60e1923d103c5fdf2dcc03770e49ec697e) - Remove one-time experiment test script (dc860cbfdaf68986a3195ecb4ecd0c00e1970469) - Remove PR description draft file (ead883563912f3fa9105b7fdfff19816330d3ea5) - Experiment 005: three arms per pair and a repeat subset (e46cb045c440a3ba4ad5fa782af984ac9440d373) - Experiment 005: strip grader-directed padding from derived answers (f8271c2267bd3fdca78346e2c6046ebe5d3d8fef) - Experiment 005: place the quoted instruction at the attack's position (c46906ab6f4d2711743ad56a026feedcfb973e39) - Experiment 005: strip the multi-turn MMR 'in the corpus' phrase (ccd35f3b87426d4396a0b3077fe1cbf2ce41560a) - Experiment 005: add Kev-4B and Laya; Modal run kit (18242937f4f1111cc800ffe62e6fc106d1b0a898) - Experiment 005: results from the 2026-10-22 H100 run (603b686b5ef754feddd4d4954b0fc8664f4dbda5) - Experiment 005: move to completed_experiments/judge_injection/experiments/005-injection-resistance (5d555f542…4cefe5d -
[experiment] Validity v1: final reference labels (reference_v1.2) (#424)
- Validity v1: final reference labels (reference_v1.2) (cca6a37f3a46c69b9be41fdac1af70344448d637)7335e7b
29 days ago
-
Expand Diana investigation with leading and adversarial framings (#421)
- Expand Diana investigation with leading and adversarial framings (e6e1a0b72c0b9a6176722f570d71ffba50e2cd54) - ground_truth_source: plain URL instead of Markdown link; adversarial difficulty 1+0 = easy (c327f9cdd712f04f8c2ea4cf91bdaad39c62a584) - ground_truth_source: restore plain URL (ec52a485ab6da1f75e10bd53965faaffe44dd41b) - difficulty rationale: match revised scores (fc41ce4a281429bc920659cd61efc4c2648fb2a3) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co>ee1d645 -
Expand Challenger investigation with leading and adversarial framings (#420)
- Expand Challenger investigation with leading and adversarial framings (c119d286a4571438f0222b788fab71b106a4f72e) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co>e2bc44b -
Add Chelyabinsk 2013 weapons-test claim question. (#418)
- Add Chelyabinsk 2013 weapons-test claim question. (8e0f281bb4cfc2cbf422eb5ee9eae98448060096) - ground_truth_source: add NASA release supporting the shock-wave clause (90e56e92741c0dc495b0d32c2e76008aeaf29101) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co>6600d2d -
Add question: Sushant Singh Rajput death Aaditya Thackeray cover up claim (#363)
- Add question: Sushant Singh Rajput death Aaditya Thackeray cover up claim (6c01d20289b35e17e117a68595720091681de68c) - Update Sushant Singh Rajput question: refined ground truth, sources, and rubric (9d4fd45555fd6b181f73cd8a3f9006439c208156) - ground_truth_answer: source the no-cover-up clause (CBI closure reports, Business Standard) (e553f2e8d7d66c3e6efed14a913ef09a0f6e8831) Co-authored-by: Jagannathan Ramanujam <jagannathanr@users.noreply.huggingface.co>eae546c