Repository navigation
Triage fixes and switch to Python 3.15 - #352
Merged
Merged
Conversation
msoedov
force-pushed
the
chore/triage-fixes
branch
2 times, most recently
from
October 18, 2026 15:53
b154b59 to
fcd586a
Compare
….9.1 Scores are identical to the 1.6.1 pickles (max abs decision diff 0.0 on 474 texts).
Relock so every dependency installs from wheels on 3.15 (multiprocess and pyyaml are pure-python sdists), allow pandas 3 and scikit-learn 1.9.
msoedov
force-pushed
the
chore/triage-fixes
branch
from
October 18, 2026 16:13
fcd586a to
32a3d23
Compare
This was referenced Oct 18, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes from the 2026-10-18 issue triage, plus the move to Python 3.15.
LLMRefusalClassifierreads the judge verdict on word boundaries, soPass.,**pass**,"pass"no longer count as refusals (which hid successful attacks). Replies with no verdict still count as refusal, as before./verifyreturns 400 for an unparseable spec and for any upstream 4xx/5xx, withUpstream returned HTTP <code>: <body>. Previously a bad spec or a 5xx target surfaced as 500.RUN_SLOW_TESTS=1). Marked 3 datasets inactive that now return 401:dayone3nder/jailbreak_prompt_JBB_sft_trainset,dayone3nder/general_safe_mix_jailbreak_prompt_JBB_trainset,ShawnMenz/DAN_jailbreak._effective_probability). Decision scores are identical to the old files on 474 texts (max abs diff 0.0).requires-python >=3.15, CI, pre-commit and docs on 3.15. Full relock: every dependency installs from wheels except multiprocess and pyyaml, which build without a C compiler. Allows pandas 3 and scikit-learn 1.9. Also clears the anyio (critical), urllib3 and mkdocs-material alerts..dockerignoreand the docker build workflow removed..flake8.Closes #350, closes #149, closes #347, closes #72, closes #135
Tests: 479 passed, 5 skipped on Python 3.15.0. New tests for #350 and #149 fail on main.
🤖 Generated with Claude Code