feat(trajectory): self-evolution line 1 — training-ready run traces#41
Merged
Conversation
Joins events.jsonl + idea tree into trajectory.jsonl (one record per idea: hypothesis, reward from eval, status, insight). Auto-written at finalize (config evolution flag export_trajectory, default on) and re-exportable for old runs via 'arbor export <s> --format trajectory'. Pure transform, no API.
…T/RL Each coordinator LLM call appends messages+output+usage to tokens.jsonl when coordinator.token_trace is on (default off, heavy). logprobs/token_ids recorded only when the provider returns them (OpenAI); Anthropic gives none, stored null rather than faked. Polar-shaped; no provider rewrite, no proxy.
Add LLMResponse.logprobs; request logprobs=true on the openai-compat chat call
and parse choice.logprobs into [{token,logprob}]. Endpoints without logprobs
(4141/Claude today) return None — interface is ready so token_trace becomes
RL-grade the moment a logprob-capable endpoint is used.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Self-evolution line 1: every run leaves training-ready data; one capture feeds SFT, RL, skill distillation, and dataset release.
trajectory.jsonl(default on): one record per idea — hypothesis, reward from eval, status, insight. Auto-written at finalize;arbor export <s> --format trajectoryfor old runs.tokens.jsonl(coordinator.token_trace, default off): per-call messages+output+usage.Pure transform, no proxy, no provider rewrite. Validated offline on a real session.