Spaces:
Running
Running
Commit History
8 days ago
-
🤖 xsft-lora results: link the LoRA adapter on the Hub and the base revision
Verified signature699a107
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] 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
-
[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
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
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 -
[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
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
-
Preserve consensus_as_of in Stage 2 question rendering (#422)
- Preserve consensus_as_of in Stage 2 question rendering (4144753e0437c650d74953021822e0e583ead76f) Co-authored-by: Chaewon Yoon <melover24@users.noreply.huggingface.co>d031397
-
Add experiment_requests/ and completed_experiments/ folders (#408)
- Add experiment_requests/ and completed_experiments/ folders (2df315e50fec0622c4d996e73c5d097524d3e9a0) - Move existing experiments into completed_experiments/, grouped by task (be58213965c5cdf5d528547d39ab6015709e18c8)4f9e6db