Milestone M4/M5, LLM-ESR official external baseline upgrade. Please inspect t…

Metadata

  • Stable ID: codex-user-prompt:d8a788278565e797
  • Source kind: codex-session-user
  • Category: coding-agent-workflow
  • Timestamp: 2026-05-09T15:12:38.205Z
  • Semantic hash: d8a788278565e797eb9278a89f73b0016a1f164e9936a94cffcc737c304cdba1
  • Public handling: selected full-text prompt with secret filtering.

Prompt Text

Milestone M4/M5, LLM-ESR official external baseline upgrade. Please inspect the local repo only and answer: what existing LLM-ESR scripts/functions can be reused to implement run-stage official support in the unified official runner? Focus on `main_train_score_llmesr_upstream_adapter.py`, `main_export_llmesr_same_candidate_task.py`, `main_generate_llmesr_sentence_embeddings.py`, and `src/baselines/official_runner/*`. Do not edit files. Report files inspected, exact functions/classes to reuse, required provenance fields, and blockers/table eligibility.

Reuse Notes

  • EXTRACTED: This is a selected Codex prompt or automation prompt from the local Codex corpus.
  • INFERRED: Future agents can reuse its structure, constraints, and acceptance criteria when creating similar Codex workflows.