Repository navigation
Conversation
A discontinuity whose crossing distance falls inside the far ladder's own range could destabilize its extrapolation and get reported as noise_dominated instead of discontinuity_suspected: the inflated far_extrapolation.error_estimate widened cross_regime's own budget enough to mask the disagreement, and far_discontinuity's check used a noise scale calibrated for the main ladder, not the far ladder's own. Fixes this by probing a noise floor local to the far ladder's own scale, scanning every adjacent-rung pair (not just the one anchored at best_index), and folding far_discontinuity.suspected into status. Also lets n_rungs_far=0 disable the far ladder (and its noise probe) entirely, for callers who'd rather trade that robustness for fewer evaluations. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This branch has not been deployed
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
far_extrapolation.error_estimateenough to widencross_regime's own budget past the disagreement it should flag, andfar_discontinuity's check used a noise scale calibrated for the main ladder rather than the far ladder's own — together, this could misreportstatus="noise_dominated"for directions that are actually near a real discontinuity.best_index), and foldingfar_discontinuity.suspectedintostatus.n_rungs_far=0as a way to disable the far ladder (and its noise probe) entirely, for callers who'd rather trade that robustness for fewer evaluations.🤖 Generated with Claude Code