kali20gakki/msAgent / skills/accuracy/rl-consistency-dump-analysis
rl-consistency-dump-analysis
Verified通用训推一致性(prefill)的 dump 数据分析 skill。分析**已采集**的训推两侧 dump:探查 dump 结构 → 语义边界映射(不要求算子同名)→ 逐边界精确门禁比对 → 定位首差异 → ULP 量化区分舍入/逻辑级 → 追放大链 → 差异点隔离验证(打桩迭代:输出打桩指令给采集侧,等它用 msprobe `load` 覆盖边界再采后比下游,枚举所有独立差异点)→ 排除根因假设 → 输出整体结论 → 产出分析报告。与具体框架/硬件解耦,适用 msprobe PrecisionDebugger 或等价工具采的任何 dump。触发词:训推一致性分析、dump 数据分析、dump 已采好帮我分析、首差异定位、所有差异点、差异点隔离、打桩指令、对齐报告、训推不一致结论、比较两侧 tensor。采集与打桩再采由 rl-prefill-consistency-dump 负责,本 skill 只消费已落盘 dump、输出打桩指令驱动再采。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 19.5 KB |
| references/analysis.md | markdown | 4.8 KB |
| references/dump-structure.md | markdown | 3.9 KB |
| references/pitfalls.md | markdown | 3.2 KB |
| references/prefill-logprob-diff.md | markdown | 4.2 KB |
| references/report-source.md | markdown | 5.3 KB |
| references/semantic-mapping.md | markdown | 5.2 KB |
| references/stubbing-loop.md | markdown | 9.0 KB |
| references/task_card.json | config | 1.6 KB |
| references/templates/alignment_report_source.zh-CN.json | config | 3.6 KB |
| references/templates/stub_instruction.example.json | config | 1.2 KB |
| scripts/analyze_dump_boundaries.py | script | 12.6 KB |
| scripts/generate_alignment_report.py | script | 87.2 KB |
| scripts/probe_dump.py | script | 9.8 KB |
references/templates/alignment_report_source.zh-CN.json
View on GitHub{
"schema_version": 1,
"language": "zh-CN",
"title": "训推 Prefill 一致性报告",
"generated_at": null,
"acceptance_contract": {
"status": "not_run",
"objective": "待填写:最终模型与目标运行路径的端到端训推一致对齐",
"requested_model_scope": "待填写",
"required_gates": [
"static_tensor_prefill",
"weight_sync",
"stateful_training",
"long_horizon_e2e",
"reduced_model_transfer",
"requested_model_final",
"production_patch_hygiene",
"clean_runtime_activation",
"performance_regression"
],
"long_horizon_required": true,
"historical_failure_step": null,
"stop_condition": "全部必过门禁都有新鲜、可复核证据"
},
"scope": {
"phase": "prefill",
"model": "待填写",
"coverage": "待填写:Prompt 长度与已验证语义边界",
"training_backend": "待填写",
"inference_backend": "待填写",
"hardware_backend": "待填写",
"evidence_provider": "待填写",
"topology": {
"tp": 1,
"pp": 1,
"ep": 1
}
},
"results": {
"native_prefill_status": "not_run",
"conditional_prefill_status": "not_run",
"unresolved_replay_ids": [],
"e2e_metric_name": "prefill_logprob_diff_mean",
"e2e_metric_status": "not_run",
"e2e_metric_value": null,
"e2e_metric_evidence": "未执行"
},
"verification_axes": {
"truth_role": "multi_axis_summary",
"tensor_prefill": {"status": "not_run", "evidence": "未执行"},
"weight_sync": {"status": "not_run", "evidence": "未执行"},
"stateful_training": {"status": "not_run", "evidence": "未执行"},
"long_horizon_e2e": {"status": "not_run", "evidence": "未执行"},
"reduced_model_transfer": {"status": "not_run", "evidence": "未执行"},
"requested_model_final": {"status": "not_run", "evidence": "未执行"},
"production_patch_hygiene": {"status": "not_run", "evidence": "未执行"},
"clean_runtime_activation": {"status": "not_run", "evidence": "未执行"},
"performance_regression": {"status": "not_run", "evidence": "未执行"}
},
"validation_matrix": {
"truth_role": "e2e_and_transfer_evidence",
"entries": []
},
"production_alignment_profile": null,
"clean_run_gate": null,
"timing": {
"session_started_at": null,
"session_ended_at": null,
"total_wall_seconds": null,
"minimum_observed_window_seconds": null,
"timing_basis": "待填写",
"module_1_seconds": null,
"module_2_seconds": null,
"module_3_seconds": null,
"module_4_seconds": null,
"module_5_seconds": null,
"module_6_seconds": null,
"discovery_seconds": null,
"repair_wave_wall_seconds": null,
"summed_repair_agent_seconds": null,
"repair_queue_wait_seconds": null,
"integration_seconds": null,
"accelerator_run_seconds": null,
"accelerator_run_count": null
},
"metric_progression": null,
"differences": [],
"comparison": {
"prompt_token_length": null,
"train_input_token_length": null,
"infer_prefill_token_length": null,
"train_prompt_slice": "待填写",
"logit_causal_shift": "待填写"
},
"provenance": {},
"artifacts": {
"case_readme": null,
"report_source": "report_source.json",
"raw_tensor_included": false,
"raw_dump_included": false,
"prompt_included": false,
"raw_training_log_included": false,
"production_patch": null,
"apply_check": null,
"production_patch_hygiene_audit": null,
"clean_runtime_audit": null,
"validation_matrix_evidence": null
},
"limitations": [
"待填写:当前证据范围与仍未解决的问题。"
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |