Repository navigation
atlas-sparkglm: pin SparkGLM RC4.2 (246a3a948ab9) - #65
Merged
Merged
Conversation
…-rc4.2-layered; image 246a3a948ab9): a repeat notice instead of masking tool calls Digest ghcr.io/enntity/atlas-sparkglm@sha256:e4d313b5e32c47545eab0059eea65a38e65d03fc8e0d5b8c2cfb497c39f0372a, install tree 246a3a948ab98d19e085bfd39797e554f0ee9261, engine dfb9c26a (tree 89e392a3) = RC4.1 + d8a55e0f. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…t 3c7d6714 Enntity/sparkglm#47 merged as 3c7d6714; its install tree is unchanged (246a3a948ab9), so the image, tag and digest stay the same. 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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Repins the
atlas-sparkglmlane from RC4.1 to SparkGLM RC4.2 (Enntity/sparkglm#47). RC4.2 is staged on both pair hosts and not yet deployed. It deploys with sparkglm-researchtools/deploy_rc42.sh.sha256:0ddea7cd…sha256:e4d313b5e32c47545eab0059eea65a38e65d03fc8e0d5b8c2cfb497c39f0372acfb463a952b5246a3a948ab9/246a3a948ab98d19e085bfd39797e554f0ee9261sparkglm/atlas-20261012-rc4.1-layered@301479a2sparkglm/atlas-20261012-rc4.2-layered@dfb9c26a(tree89e392a3)401a5c2793d220b0(branchrelease/rc4.2)What changes for clients. After the same tool call returns the same non-error result twice, the engine appends a
<REPEAT NOTICE>to that tool result and thinks atxhighon that turn. It no longer hard-masks<tool_call>, which had left GLM narrating its intended action as prose (756 tokens in the production incident, about 4,000 on the reproduction). The profiles and the decode path are unchanged.Files:
pins.json, the recipe, the backend README,THIRD_PARTY_NOTICES.md,docs/recipes.md, and the recipe test. The changes are the identity substitutions made by sparkglm-researchtools/lloom_pin_rc42.py.Checks:
node test/atlas-sparkglm-recipe.test.mjspasses.node backends/atlas-sparkglm/verify-pins.mjspasses.node test/smoke.mjspasses.Before merging:
93d220b058921f83c253a3a5544c44f141490989(and93d220b0) with the merge commit in the files above. The tree, tag and digest stay the same.🤖 Generated with Claude Code