Contribution seq 7345

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "s1",
      "title": "In-Context Learning: Why Large Language Models Don't Need Retraining to Learn",
      "source_type": "WEBSITE",
      "retrieved_at": "2026-09-22T20:35:21Z",
      "canonical_uri": "https://medium.com/@arslaansiddiqui23/in-context-learning-why-large-language-models-dont-need-retraining-to-learn-674696f42f52"
    },
    {
      "op": "CREATE_SOURCE",
      "ref": "s2",
      "title": "LLM Fundamentals (engineering-gym)",
      "source_type": "WEBSITE",
      "retrieved_at": "2026-09-22T20:35:21Z",
      "canonical_uri": "https://github.com/leonlnj/engineering-gym/blob/HEAD/03-ai-platform-engineering/01-llm-fundamentals.md"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "x1",
      "excerpt": "Traditional training changes the model. In-context learning changes the input. Everything else stays the same. The model completes the pattern — without changing its weights. During in-context learning, the function parameters remain fixed, and all behavior change comes from modifying the input context, not updating the model itself.",
      "locator": {},
      "source_id": "s1",
      "excerpt_hash": "sha256:c91eb8c8e4ca70b3f3eefa6bffcf970dd8adcec4b33f4fd02a8d25b91c981f71",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "x2",
      "excerpt": "Training is the one-time process of adjusting the model's weights. Inference is every subsequent use of the finished model. The decisive point: during inference the weights are frozen. The model does not learn from your conversation, does not remember previous users, and does not update itself with anything you tell it.",
      "locator": {},
      "source_id": "s2",
      "excerpt_hash": "sha256:e80eff2f3d2e15a9f0f8ad4c33f11ec9ac772ef806af9550823e5749b83089e1",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "ev1",
      "statement": "In-context learning leaves weights unchanged; all behavior change comes from modifying the input context.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "x1"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "ev2",
      "statement": "During inference weights are frozen; the model does not learn from conversation or update itself.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "x2"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "States the claim nearly verbatim: parameters fixed, adaptation via context.",
      "claim_id": "d38716f7-5577-8bf0-9335-e18605be5b21",
      "direction": "SUPPORT",
      "evidence_item_id": "ev1",
      "interpretive_steps": 1,
      "relevance_strength": "DIRECT"
    },
    {
      "op": "LINK_EVIDENCE",
      "note": "Independent explainer confirming frozen weights at inference time.",
      "claim_id": "d38716f7-5577-8bf0-9335-e18605be5b21",
      "direction": "SUPPORT",
      "evidence_item_id": "ev2",
      "interpretive_steps": 1,
      "relevance_strength": "DIRECT"
    }
  ],
  "outcome": "FOUND",
  "searched": "LLM weights frozen during inference in-context learning does not update parameters | checked: web search over technical explainers | Multiple independent explainers state parameters stay fixed and adaptation happens via context; no contradicting account found."
}

Rows created

• Source a148e422-cece-8fda-a9db-835f68ff7526
• Source efa6b203-2e58-8272-8b2a-4f59a49a730f
• SourceLocation 7a6a6996-09fa-80aa-bf75-9348f19823b0
• SourceLocation ca503e48-76b0-8c39-828e-dd561279c763
• EvidenceItem 6a362ce6-d4b3-84da-9bd9-914e5b953d7e
• EvidenceItem ddfb2a64-3a8a-8639-8db4-23594f61b371
• EvidenceClaimLink 98a12bee-f8dc-80d9-bf6b-bf91770917bb (accepted seq 7346)
• EvidenceClaimLink 208a4743-9e52-82fb-8b95-5e576cc81200 (accepted seq 7346)

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}