Repository navigation
fix(claims): flag only asserted claims in replies; keep honest rows in training - #9
Merged
Conversation
…n training flagged_in_reply() now checks only register entries with explicit `asserts` patterns (currently OSIRIS_ON_QPU), and a sentence that reports a claim as refuted, failed or untrue is not an assertion. Previously, naming a topic (e.g. "retiring constants like theta_lock") flagged the reply and dropped it from training. Ledger dry run (~/.osiris/living/exchanges.jsonl): 921 kept, 1 left out (a test exchange). Full suite: 1803 passed, 2 failed; both failures pre-exist on main (test_exec_policy symlink case; GitHubBridge auth test expects the old osiris-dnalang account). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What
flagged_in_reply()checks only register entries that carry explicitassertspatterns (currentlyOSIRIS_ON_QPU). A sentence that reports a claim as refuted, failed or untrue does not count as asserting it.train.chat_lessonsleaves out only replies that are actually flagged.Why
Before this change, a reply that merely named a topic (for example "retiring constants like θ_lock") was flagged and dropped from training, so honest replies were being discarded.
Evidence
test_claims,test_train,test_living: 204 passed.mainwithout this change:test_exec_policy::test_symlink_named_ls_pointing_at_rm_is_treated_as_rmTestGitHubBridge::test_auth_status_inspection(expects the oldosiris-dnalangaccount)ruff check --select E,F,Won the changed files: clean.Caveat
Assertion detection is a word test, not semantic reading. Quoted text and hypotheticals can still be misclassified.
🤖 Generated with Claude Code