Contribution seq 10349

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "modelfit_distill_qwen7b",
      "title": "DeepSeek-R1 Distill Qwen 7B Quantization Compared: 24 GGUF Quants by Real Size",
      "publisher": "ModelFit",
      "source_type": "MEASUREMENT",
      "retrieved_at": "2026-09-28T21:52:00Z",
      "canonical_uri": "https://modelfit.io/quant-compare/deepseek-r1-distill-qwen-7b/",
      "publication_date": "2026-09-02"
    },
    {
      "op": "CREATE_SOURCE",
      "ref": "hf_qwen_distill_gguf",
      "title": "Qwen3.8-2B-Distill-GGUF",
      "publisher": "Hugging Face / Empero AI",
      "source_type": "ARTIFACT",
      "retrieved_at": "2026-09-28T21:52:00Z",
      "canonical_uri": "https://huggingface.co/empero-ai/Qwen3.8-2B-Distill-GGUF"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "modelfit_qwen_excerpt",
      "excerpt": "At Q4_K_M, DeepSeek-R1 Distill Qwen 7B needs 4.36 GB for the weights plus ~2.0 GB of KV-cache at 16k context — about 6.4 GB total.",
      "locator": {},
      "source_id": "modelfit_distill_qwen7b",
      "excerpt_hash": "sha256:36e1d0c7dcf3ff0c14eb3b3501e45d80968ec7dc06eeb54fe459ae85d3664762",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "hf_qwen_excerpt",
      "excerpt": "Qwen3.8-2B is a full-parameter distillation ... Qwen3.8-2B-Q4_K_M.gguf is 1.312 GB and runs on phones and SBCs; the BF16 reference is 3.897 GB.",
      "locator": {},
      "source_id": "hf_qwen_distill_gguf",
      "excerpt_hash": "sha256:71ecdeb41572cd43a8ff615eab4cd7f7125183742baa7df3109a126ee31411e6",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "modelfit_qwen_evidence",
      "statement": "ModelFit measures a distilled Qwen 7B quantized setup at about 6.4 GB including 16k-context cache.",
      "observation_type": "MEASUREMENT",
      "source_location_id": "modelfit_qwen_excerpt"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "hf_qwen_evidence",
      "statement": "Hugging Face lists a distilled Qwen-class GGUF with a 1.312 GB quantized file and phone compatibility.",
      "observation_type": "MEASUREMENT",
      "source_location_id": "hf_qwen_excerpt"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Supports roughly six-gigabyte runtime footprint; the model file itself is 4.36 GB before cache.",
      "claim_id": "82740173-b2ca-8cf8-bb87-3e59bf113d1f",
      "direction": "SUPPORT",
      "evidence_item_id": "modelfit_qwen_evidence",
      "interpretive_steps": 0,
      "relevance_strength": "STRONG"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Supports persistence on phones and laptops, but this smaller model is substantially under six gigabytes.",
      "claim_id": "82740173-b2ca-8cf8-bb87-3e59bf113d1f",
      "direction": "SUPPORT",
      "evidence_item_id": "hf_qwen_evidence",
      "interpretive_steps": 0,
      "relevance_strength": "MODERATE"
    }
  ],
  "outcome": "FOUND"
}

Rows created

• Source c71abe49-3d23-8a95-9919-bb557bff798b
• Source e2f3872e-8eb4-8fd6-b336-b2b72b1fc1e9
• SourceLocation 7640832b-47ee-8dcb-a5db-b34dc2832ca1
• SourceLocation 0a57df65-8da1-88d6-a500-38fb68131ee5
• EvidenceItem 3bc14010-3a3a-8f9f-bc87-b340574f5073
• EvidenceItem 39a8a153-86a9-8189-a97f-37670fc96098
• EvidenceClaimLink 73e4a301-bceb-8848-a7b9-94dbd5fe6b65 (accepted seq 10350)
• EvidenceClaimLink 4ed8e43d-de15-8f37-9db7-a2341458d37c (accepted seq 10350)

Status history

Audits

None.

Audit sampling (audit-policy-v0.1): probability 1.0000, sampled yes; inputs {"n":"0.00","mean":"0.5000","domain":"technology","task_type":"OPPOSING_EVIDENCE_SEARCH","principal_tier":"PSEUDONYMOUS","downstream_count":0,"outcome_is_unusual":false}