AGENT SKILL

pplm_attribute_models

Use when a recovery or implementation needs differentiable PPLM attribute scores for BoW or simple discriminator-style control.

License 许可
Apache-2.0
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
pplm_attribute_models
description
Use when a recovery or implementation needs differentiable PPLM attribute scores for BoW or simple discriminator-style control.

02 资源文件 2

03 SKILL.md 正文

PPLM Attribute Models

Use this skill to build the plug-in attribute model p(a|x) used by PPLM. Do not use it to modify base LM weights.

Inputs

  • Vocabulary tokens.
  • Topic BoW terms or linear classifier weights.
  • Candidate token probabilities or a hidden vector.

Outputs

  • Numeric attribute score.
  • Attribute loss suitable for gradient ascent/descent in a controller.
  • Diagnostics for missing BoW terms.

Workflow

  1. Normalize vocabulary and target words.
  2. For BoW control, score the probability mass assigned to target words using -log(sum p(target)) as a minimization loss.
  3. For discriminator-style control, apply a linear score or softmax outside the base LM.
  4. Return diagnostics; never update the base LM.

Validation

Run python tests/test_attribute_models.py.

Limitations

Reduced proxy implementations must declare that they are not full GPT-2 345M reproduction.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351