fix(bench): WORKER_TEMPERATURE knob — kimi-class models 400 on temperature ≠ 1#260
Merged
Conversation
…ature ≠ 1 The kimi matrix row's zeros were a broken channel both times: the model rejects runShot's hardcoded temperature 0.7 with a 400 (probe-verified: 'only 1 is allowed for this model'), downing every rollout. The temperature is now per-run settable on all three runners; kimi re-runs with WORKER_TEMPERATURE=1.
tangletools
approved these changes
Jun 11, 2026
tangletools
left a comment
tangletools
left a comment
Contributor
There was a problem hiding this comment.
✅ Auto-approved PR — 52405390
Blanket team auto-approval is enabled for this reviewer service.
The full PR reviewer audit still runs separately and will publish findings if it detects issues.
tangletools · auto-approval · reason: blanket_auto_approve · 2026-06-11T12:35:48Z
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Second kimi autopsy, definitive root cause: the model rejects any
temperature ≠ 1— probe with runShot's exact request shape returns 400\"invalid temperature: only 1 is allowed for this model\". Every rollout died on the 400 (every rollout went downin the per-strategy errors), so both kimi matrix rows measured a broken channel, not the model — the earlier maxTokens hypothesis was necessary but not sufficient.WORKER_TEMPERATUREenv on all three runners; kimi re-runs at 1.0. Bench typecheck ✓.