SA260324R-0129
① SA Source
- Source: 開啟完整 SA 文章
- Section:
Speculative Decoding - Line hint:
167
Context Before
For LPUs, deploying a draft model or MTP layers is quite different from applying AFD.
Evidence
FFNs are stateless, while draft models
Context After
and MTP layers require dynamic KV cache loading. Each FFN is around hundreds of megabytes, whereas draft models and MTP layers take up tens of gigabytes. To support this memory usage, LPUs can access up to 256 GB of DDR5 per Fabric Expansion Logic FPGAs on the LPX compute tray.
② Atomic Claim
FFN 是 stateless，而 draft model……
- Epistemic Mode:
ASSERTED - Mapping Status:
PARTIAL
③ Semantic Frame
{
"frame_type": "CLAIM_ONLY",
"qualifiers": {
"condition_text": "原子事實在 draft models 處中斷，不能自行補完後半句",
"numeric_mentions": [],
"temporal_mentions": []
},
"reason": "EXTRACTION_FRAGMENT_INCOMPLETE_AFTER_DRAFT_MODELS",
"statement": "FFN 是 stateless，而 draft model……"
}④ Canonical Entity Mapping
| Role | Surface Label | Canonical Target |
|---|---|---|
| — | — | — |
⑤ Human Review
請在 Properties 逐項確認：
- 原文 → Atomic Claim 是否忠實
- Atomic Claim → Semantic Frame 是否忠實
- Canonical Entity mapping 是否正確
- Epistemic mode 是否保留原文語氣
- 最後選擇
review_action
Review state
Markdown 內文不是正式 approval。只有 Apply bridge 寫入的 Decision Ledger event 才是正式決策。