clm_52a443f3a2e54c56a690a701e62046ca_r1

IX2-0199

① SA Source

• Source: 開啟完整 SA 文章
• Section: Unpacking Inference Providers’ Unit Economics
• Line hint: 352

Context Before

Source: SemiAnalysis InferenceX

At this high of interactivity, it becomes necessary to employ speculative decoding techniques like MTP to achieve high enough throughput to make inference economical. Luckily, MTP can increase throughput with relatively low risk to overall model accuracy. We will go on to talk more about MTP, and how it can be applied to increase throughput / decrease cost, in later sections of this article.

Evidence

Lastly, we show one more chart of an FP8 DeepSeek workload served at 125 tok/s/user

Context After

Source: SemiAnalysis InferenceX ↗

② Atomic Claim

最後，文章再展示一張 FP8 DeepSeek workload 在 125 tok/s/user 下的圖表。

• Epistemic Mode: ASSERTED
• Mapping Status: COMPLETE

③ Semantic Frame

{
  "attribute": "THROUGHPUT",
  "context_nodes": [
    {
      "id": "02_companies/DeepSeek",
      "label": "DeepSeek"
    }
  ],
  "entity": {
    "id": "04_knowledge_base/FP8",
    "label": "FP8"
  },
  "frame_type": "ATTRIBUTE",
  "qualifiers": {
    "condition_text": null,
    "numeric_mentions": [
      "125 tok/s/user"
    ],
    "temporal_mentions": []
  },
  "value": {
    "numeric_mentions": [
      "125 tok/s/user"
    ],
    "value_text": "最後，文章再展示一張 FP8 DeepSeek workload 在 125 tok/s/user 下的圖表。"
  }
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
entity	FP8	FP8
context_0	DeepSeek	DeepSeek

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。