cad_recovery_evaluation
Evaluate CAD conflict-recovery runs with exact-match metrics and mechanism-check summaries.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- cad_recovery_evaluation
- description
- Evaluate CAD conflict-recovery runs with exact-match metrics and mechanism-check summaries.
02 资源文件 2
03 SKILL.md 正文
CAD Recovery Evaluation
Use this skill when assessing whether a CAD experiment followed supplied context rather than a memorized prior in QA-style conflict items.
Inputs
- Items containing
id,context_answer, andprior_answer. - Regular decoding predictions.
- CAD predictions.
- Per-item traces showing prompt separation, dual logits, logit adjustment, and token selection.
Outputs
Exact-match metrics, improvement, per-item records, and mechanism-check booleans suitable for recovery_result.json.
Workflow
- Normalize answers by lowercasing and stripping whitespace/punctuation.
- Score regular and CAD predictions against the context answer.
- Verify every item has prompt separation, dual logits, adjusted logits, and selection trace.
- Return metric and mechanism summaries.
Validation
Run tests/test_recovery_evaluation.py or Distiller validation.
Limitations
This evaluator is for reduced conflict QA recovery, not ROUGE or factuality metrics on summarization.
04 同仓库更多 Skills
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.
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.
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.
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.
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.
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).