kali20gakki/msAgent / skills/accuracy/rl-prefill-consistency-dump
rl-prefill-consistency-dump
Verified训推一致性(prefill)采集侧:为训练/推理两侧采集可比的 prefill dump(输入对齐 PROMPTS_ONLY → 参数对齐校验(dump 前前置:校验两侧 config 超参/结构/权重,不一致输出参数不一致点并对齐)→ msprobe 接入采集 → 采集验证 → 恢复代码)。按用户需求判断用法:① 仅采集 dump——用户只要 dump 数据,走采集主流程(0→7→9),直接产出两侧可比 dump;② 配合分析——用户要求找所有独立差异点/差异点隔离验证/配合 rl-consistency-dump-analysis 分析时,才按分析侧打桩指令用 msprobe `load` 覆盖边界二次采集(阶段 8)再恢复代码。差异分析、首差异定位、打桩指令、整体结论由 rl-consistency-dump-analysis 承担。触发词:训推一致、训推一致性、prefill 比对、msprobe dump/采集、精度比对数据采集、打桩二次采集、参数对齐校验、参数不一致点、mHC 配置不一致、权重不同步、清卡验环境后采数据;已采 dump 的分析转 rl-consistency-dump-analysis。
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 25.7 KB |
| references/msprobe-config.md | markdown | 5.4 KB |
| references/new-framework.md | markdown | 5.2 KB |
| references/param-alignment.md | markdown | 14.4 KB |
| references/pitfalls.md | markdown | 22.1 KB |
| references/prefill-logprob-diff.md | markdown | 5.6 KB |
| references/probs-diff-monitoring.md | markdown | 6.6 KB |
| references/slime.md | markdown | 5.4 KB |
| references/stubbing.md | markdown | 7.7 KB |
| references/task_card.json | config | 3.2 KB |
| references/templates/param_alignment_instruction.example.json | config | 1.4 KB |
| references/templates/stub_instruction.example.json | config | 1.2 KB |
| references/verl.md | markdown | 5.0 KB |
| references/vime.md | markdown | 9.5 KB |
| scripts/param_align_check.py | script | 8.7 KB |
| scripts/param_align_provider_dump.py | script | 5.7 KB |
| scripts/param_alignment_mapping.example.json | config | 2.3 KB |
references/templates/param_alignment_instruction.example.json
View on GitHub{
"round": 0,
"mismatch_points": ["P1", "P2"],
"alignment": [
{
"point_id": "P1",
"align_side": "train",
"semantic": "Hyper-Connections (mHC) enable 开关",
"method": "config/code",
"target": {"field": "enable_mhc_connections", "value": true},
"action": "修 Megatron-Bridge deepseek_v4_bridge.py: L463 enable_hyper_connections -> enable_mhc_connections (上游字段名笔误, 训练侧开关落空); 重建训练模型后重跑校验",
"verify": "重跑参数对齐校验, P1 变 EXACT; 训练侧模型结构出现 decoder.hc_head_* / layers.*.hc_attn_* 参数"
},
{
"point_id": "P2",
"align_side": "train",
"semantic": "Hyper-Connections 流数 (hc_mult)",
"method": "config/code",
"target": {"field": "mhc_num_residual_streams", "value": 4},
"action": "deepseek_v4_bridge.py L465 num_residual_streams -> mhc_num_residual_streams; 与 config.json hc_mult=4 一致",
"verify": "重跑校验, P2 变 EXACT"
}
],
"then": "对齐 -> 重建模型 -> 重跑校验(阶段 5)全部 EXACT -> 接入采集(阶段 6)跑 dump; 对齐改动登记进改动清单, 阶段 9 一并恢复",
"note": "对齐方向以权威侧为准: 推理侧 vllm 直接读 config.json 启用 mHC (权威), 训练侧对齐到推理侧. msprobe load 不覆盖参数, 本指令走 config/code 重建, 不走 load"
}
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 |