AGENT SKILL

cad_logit_adjustment

Apply the Context-aware Decoding contrastive logit formula and report mechanism diagnostics for recovery experiments.

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

01 前言 Front Matter

name
cad_logit_adjustment
description
Apply the Context-aware Decoding contrastive logit formula and report mechanism diagnostics for recovery experiments.

02 资源文件 2

03 SKILL.md 正文

CAD Logit Adjustment

Use this skill for the paper's core next-token transformation: adjusted = (1 + alpha) * logits(context, query, prefix) - alpha * logits(query, prefix). It does not construct prompts or score final answers.

Inputs

  • Context-conditioned logits as a token-to-score mapping.
  • Query-only/prior logits as a token-to-score mapping.
  • alpha, usually 0.5 for summarization or 1.0 for knowledge-conflict tasks.

Outputs

Adjusted logits over the shared vocabulary plus diagnostics: alpha, regular argmax, prior argmax, adjusted argmax, and whether CAD changed the selected token.

Workflow

  1. Verify identical vocabularies.
  2. Compute the CAD formula for each token.
  3. Preserve logits without softmax unless probabilities are requested downstream.
  4. Report argmax diagnostics for mechanism checks.

Validation

Run tests/test_logit_adjustment.py or Distiller validation.

Limitations

This script works on one decoding step; generation loops should call it repeatedly with updated prefixes.

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