mirror of
https://github.com/wassname/jsteer.git
synced 2026-10-28 11:50:16 +08:00
wassnameandClaudypoo
1f458fd53e
Ported from the verified j-steer-dev experiment (word pullback beat random on 3/5 moral foundations, Qwen3-4B n=3). jacobian.py wraps jlens fit/save/ load and derives steering vectors by CPU matvec against the cached J; vjp.py is the one-backward parity path; applies.py registers the methods and delivery modes into steering-lite. Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
2.2 KiB
2.2 KiB
jsteer -- agent notes
Plan of record: /home/wassname/.claude/plans/review-specs-00-minimal-experiment-md-an-peppy-sky.md Evidence base: ../j-steer-dev/docs/RESEARCH_JOURNAL.md (verified 3/5 word-steering result).
What this is
repeng-style UX for Jacobian pullback steering. Core algo files are WRITTEN (by the main agent, ported from verified j-steer-dev code -- do not rewrite the math, it is parity-gated against the verified experiment):
jsteer/jacobian.py-- Jacobian.fit/save/load (wraps jlens, the researchers' verified primary code) + word/persona/persona_topk/random vectors -> steering_lite.Vector.jsteer/applies.py-- steering-lite method registration + delivery modes.jsteer/vjp.py-- direct VJP path; the parity reference for the cache.
Runtime is steering-lite: with v(model, C=8): model.generate(...).
Remaining work (task list has details; U-numbers from the plan)
- uv scaffold:
uv sync(torch cu121+ index if needed), fix any import errors fail-fast (no defensive fallbacks). LOCAL DEV: you may switch [tool.uv.sources] to path deps (../j-steer-dev/docs/vendor/jacobian-lens and ../../lite/steering-lite, editable) if the git fetches are slow -- leave a comment saying which is active and why. - Smoke on Qwen/Qwen3-0.6B: tiny fit (8 short web-text prompts, mid layers, dim_batch 8), word_vector(["happy","joy"]), generate at C in {-8, 0, 8}, print FULL first prompt + generations (token-efficient-logging skill).
- U1 parity gate BEFORE demos: cos(Jacobian-cache pullback, word_vector_vjp) per layer > 0.999, same prompts/max_length/skip_first. If it fails, that is a bug in the wiring (the math is linear-identical), debug do not tune.
- 0.6B real fit (~64 prompts) cached to artifacts/; 4B via pueue (label why:/resolve:).
- notebooks/word_steering.ipynb (hello-world), persona_steering.ipynb (persona variants are EXPERIMENTAL -- they failed specificity controls in j-steer-dev; keep that framing), lens_readout.ipynb optional.
- README: classic-repeng length, honest evidence section.
Style
Fail fast, no defensive programming, loguru, no LLM-tell prose in README. Comments marked as Claude-authored where opinionated.