Repository navigation
docs(eval): frontier reasoner detects natural-language supersessions (bi-temporal detection resolved) - #120
Conversation
…(bi-temporal detection resolved) Direct resolve-prompt probe vs gpt-5.6-sol (OAuth, concurrent): natural-language knowledge-updates (5K PB, Berlin->Munich) correctly return 'supersede', control returns 'insert'. The heuristic fired 0x; frontier detects + discriminates. Closes the bi-temporal root question: blocker was detection quality, not mechanism or answerer. Enabled by the async write path + a fast concurrent frontier backend. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Your trial has ended. Reactivate Greptile to resume code reviews.
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: d08e77c17f
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| unblocked it: (1) the async write path (enrichment off the critical path, concurrent), and (2) a fast, | ||
| **concurrent, frontier** LLM backend — the ChatGPT/OpenAI OAuth `gpt-5.6-sol` served directly over its | ||
| responses endpoint (8 parallel calls in 4.1 s, vs the codex CLI's ~15 s serial process-spawn). |
There was a problem hiding this comment.
The document’s stated provenance says every number comes from a measured run_eval invocation, but this new timing is from a direct OAuth Responses-endpoint probe. run_eval only configures the library reasoner via SYNAPTIC_LLM_URL/SYNAPTIC_LLM_MODEL and ingests dataset conversations; it has no mode to issue this direct four-case resolve probe or report its parallel-call timing. As written, readers will incorrectly treat the 4.1 s / 15 s measurements and the four verdicts as reproducible evaluation-harness results; add the probe command/method and explicitly exempt it from the harness provenance claim (or move it to a separately labeled experiment).
Useful? React with 👍 / 👎.
Closes the bi-temporal thread's root question. Earlier: the heuristic reasoner fired 0 supersessions on natural-language knowledge-updates (weak-answerer run: 40/40 identical predictions), and an LLM-reasoner run was aborted as impractical (~4h serial).
Two unblocks: (1) the async write path (#117, enrichment off the critical path, concurrent), and (2) a fast concurrent frontier backend — the OAuth
gpt-5.6-solserved directly over its responses endpoint (8 parallel calls in 4.1s vs the codex CLI's ~15s serial spawn).Direct probe of the library's exact
resolveprompt against the frontier model:The frontier reasoner detects natural-language supersessions and discriminates (control → insert) — what the heuristic could not. So bi-temporal's blocker was detection quality, not the retrieval mechanism (unit-tested) or answerer masking. The end-to-end value A/B (does dropping these detected stale facts help a weak answerer) is the natural next step, now feasible on the async path + concurrent backend.
Docs-only.
🤖 Generated with Claude Code