Dashboard › institutional-transition-lab › Distillation
433c03b0-6beb-4498-bd61-4ecb2f25996f["lore_tm_v1_lfGcSBIp7volGbOu-u6ohWJ2MjU90W9FCtuxeBstHV8","lore_tm_v1_dKH8plXYOQsAoGQQ1xsiMuZkKYmOd_fFJ31NC4v4b9Y","lore_tm_v1_rw2f3C-vek8rli5TeNDIR3hfscgBQE9RjZTBnD-YGb8","lore_tm_v1_5ezyB6oKC2pNgKQYlK2i4DeQ95v6muOGfjoM2h98tjs","lore_tm_v1_cBnkbzaA4vHtL2R504CpqfUU3827naaTLoempHbV45M","lore_tm_v1_jhlHsfp60IhKjR0LVAGMrHLLBVgt3tkTZcGgmY--kEE","lore_tm_v1_RNjylN28_yrI9MYQTRne1aupV5QUaM4OUfJ5g-OYJck"]
/home/byk/Code/institutional-transition-lab/schema/governance-coding-v1.schema.json, a JSON Schema Draft 2020-12 object titled Performance-blinded governance coding v1; it sets additionalProperties: false and requires record_class, event_kind, title, announced_on, effective_on, affected_scope, body_patch_relation, power_changes, confidence, and ambiguity.record_class enum in exact order: effective_institutional_change, announced_institutional_change, proposal_only, control_event, no_event, abstain; and event_kind: leadership, board_or_steering, control_rights, reorganization, foundation_transfer, license, fork, reunification, strategy, product, external, or null.body_patch_relation enum in exact order: consistent, patch_supersedes_body, body_only, patch_only, conflict, not_applicable, unclear; confidence must be a number from 0 through 1; announced_on and effective_on are nullable date-formatted strings.power_changes item disallows additional properties and requires actor, right_kind, target, direction, change_status, scope, and evidence_refs; actor, target, and scope are nonempty strings.right_kind enum: appoint, remove, elect, vote, delegate, override, approve, merge, write, release, veto, own, license, steward, fund, set_budget, set_strategy, set_policy, set_membership, set_terms, inform; direction is added, removed, or modified; change_status is effective, announced, proposed, rejected, or unclear.power_changes[].evidence_refs array must contain at least 1 item, and each item permits only the required nonempty string fields source_id and quote.jq: Unknown option --argfile; jq advised using jq --help or the online documentation at https://jqlang.github.io/jq.body, pr_metadata, and patch:0 through patch:11. body alone had exact_body_quote: true; every source had exact_single_line_quote: false; patch:6 alone had policy_quote: true; patch:1 alone had autonomy_quote: true; all other displayed flags were false./home/byk/Code/institutional-transition-lab/src/institution_lab/llm_event_probe.py lines 482, 566, and 592 use quote_count, grounded_quote_count, substring grounding via quote in str(document["text"]), and _safe_ratio(grounded_quote_count, quote_count); /home/byk/Code/institutional-transition-lab/src/institution_lab/governance_coding.py line 302 emits "{prefix}.evidence_refs[{ref_index}] is not grounded"; /home/byk/Code/institutional-transition-lab/src/institution_lab/llm_candidate_triage.py lines 105–106, 157, 162, 181, and 183 track raw and processed quote grounding and calculate rates./home/byk/Code/institutional-transition-lab/src/institution_lab/governance_coding.py lines 260–315: confidence rejects booleans, non-numeric values, non-finite values, and values outside [0,1], emitting confidence must be between zero and one.governance_coding.py: source_map maps each source’s source_id to text; power_changes must be an array; every change must have exactly POWER_KEYS; actor, target, and scope must be nonempty; and right_kind, direction, and change_status must belong to RIGHT_KINDS, DIRECTIONS, and CHANGE_STATUSES.governance_coding.py: evidence_refs must be a nonempty list; each reference must have exactly REF_KEYS; and a reference is rejected as not grounded if its source text is absent/non-string, its quote is empty, or the quote is not an exact substring of the corresponding source text.governance_coding.py: effective_institutional_change, announced_institutional_change, and proposal_only require nonempty power_changes; all other classes must not contain power_changes; abstain requires non-null ambiguity./home/byk/Code/institutional-transition-lab/tests/test_governance_adjudication.py found 7 matches involving accepted_from at lines 106 and 150, graph_edge_revision at lines 113, 153, and 213, and decision: "revise" plus accepted_from: None at lines 192–193.workflow_run_id: 33063470968, artifact_id: 9642983578, artifact_sha256: "artifact-hash", documents_sha256: "filled-by-test", documents_file_sha256: "documents-file-hash", coding_package_file_sha256: "coding-package-file-hash", protocol_sha256: "protocol-hash", coding_schema_sha256: "coding-schema-hash", and adjudication_schema_sha256: "adjudication-schema-hash".reviewer_id: "opencode-gpt-5.6-sol-primary", reviewer_type: "llm", model: "openai/gpt-5.6-sol", role: "source_adjudicator", and assistance Primary source review and synthesis.record_id: "example-1" and source URL https://github.com/acme/example/issues/1: decision accept, accepted_from: "luna_a", reviewer opencode-gpt-5.6-sol-primary, rationale The source contains no institutional transition., null evidence_limitations and unresolved_reason, empty claim_evidence_refs, and graph_edge_revision with basis: "luna_a" plus empty added and removed arrays._pins() in /home/byk/Code/institutional-transition-lab/tests/test_governance_adjudication.py, returning the six exact hashes artifact-hash, documents-file-hash, coding-package-file-hash, protocol-hash, coding-schema-hash, and adjudication-schema-hash under their corresponding SHA-256 keys._validate() test behavior: it imports canonical_sha256, hashes _documents(), writes that value to both package["summary"]["documents_sha256"] and adjudication["inputs"]["documents_sha256"], then calls validate_adjudication(adjudication, documents, package, verified_pins=_pins()).test_one_abstention_keeps_canonicalization_gate_closed: changing the sole record to decision: "abstain", accepted_from: None, unresolved reason Required source evidence is incomplete., final_coding: None, and graph_edge_revision: None yields no errors, record_count == 1, unresolved_record_ids == ["example-1"], and canonicalization_gate == "closed".test_record_ids_must_match_frozen_documents_exactly_once with mutations duplicate and missing; either duplicating the record or clearing the records produces an error containing record IDs must match frozen documents exactly once and keeps canonicalization_gate == "closed".test_accept_must_equal_the_named_frozen_response: changing accepted final_coding.confidence to 0.5 produces an error containing accepted coding does not equal luna_a and keeps canonicalization_gate == "closed".test_institutional_claims_require_grounded_claim_and_edge_evidence: a revised effective_institutional_change with event_kind: "control_rights", title Council gains release approval, dates 2024-04-01, affected scope releases, body_patch_relation: "body_only", and invented body quotes in both claim and edge evidence produces a not grounded error and keeps canonicalization_gate == "closed".The council, right_kind: "approve", target releases, direction added, status effective, scope project releases, and evidence reference {"source_id": "body", "quote": "invented"}; its graph_edge_revision uses basis: "luna_a" with empty added and removed.test_gemini_cannot_adjudicate_or_assist: changing the reviewer model to google/gemini-3.7-flash produces an error containing Gemini is restricted to event triage and keeps canonicalization_gate == "closed".