The full six-domain battery in one run: 53/58 - #40
Merged
Conversation
One sitting, one model, 62 questions -- 58 gradeable across arithmetic, units, logic, physics and calculus, plus 4 consistency probes reported apart: arithmetic 18/20 logic 8/8 physics 14/14 calculus 5/8 units 8/8 consistency: anchored 2/2 steady, invented 0/2 steady -- both as designed 53/58 (91%), zero errored runs, pinned. Calculus is the weak domain and its misses are legible model errors: a dropped chain factor (6912 for 235824), an integral off by 10, a tangent slope read at the wrong point. The consistency probe again held its shape: steady on anchored facts, unsteady on invented places, and never counted as correctness.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
One sitting, one model (qwen3:4b-instruct-2507), 62 questions — 58 gradeable plus 4 consistency probes reported apart.
53/58 (91%), zero errored runs, pinned as the number the README's positioning rests on.
Calculus is the weak domain, and its misses are legible model errors — a dropped chain factor (6912 for 235824), an integral off by 10, a tangent slope evaluated at the wrong point. Physics at 14/14 across seven groups is the strongest single-domain result the battery has produced.
No code changes; one claims file and two pinned numbers.