experiment: audit graph provenance

This commit is contained in:
2026-08-07 13:01:25 +01:00
parent d0b9be5fe6
commit 5892a9b0f6
2 changed files with 102 additions and 3 deletions
+3 -3
View File
@@ -124,8 +124,8 @@ Answer before continuing:
---
*Created by Experiment 34. Updated by Experiments 3852I and 53. Branch: `feature/user-workspace-ux-v0.7`.*
*Created by Experiment 34. Updated by Experiments 3853, 54A. Branch: `feature/user-workspace-ux-v0.7`.*
### Return-to-Work Note (Experiment 52I)
### Return-to-Work Note (Experiment 53 → 54A)
Experiments 52F52H established that the model strengthens vague relevance phrasing beyond what was supplied: four of five ambiguous statements became `could_change_decision`, only "connected to" preserved `cannot_determine`. Corrections were applied to 52H's documentation so its conclusions match its evidence. Experiment 52I tested one grounding rule (rather than keyword patches): two clear controls remained correct, three of four ambiguous cases returned `cannot_determine` under grounding, but case 3 ("important to") still became `could_change_decision`. The remaining defect is primarily grounding — the model can still add relationship strength that the supplied meaning did not establish. Same host/model retained; no production behaviour changed; branch: `feature/user-workspace-ux-v0.7` (commit pending). First file to inspect when resuming: `docs/design-evolution-log.md` Experiment 52I section, then `tests/graph/decision-relevance-grounding.test.js`.
Experiment 53 proved semantic separation was possible; 54A inspected whether the current SituationGraph can preserve that distinction from source code alone. The audit confirms the graph's kind/status/confidence fields describe epistemic status only — there is no provenance field recording where information came from. Supplied-versus-inferred provenance is not recoverable from validated graph state. No production code changed; branch: `feature/user-workspace-ux-v0.7`. What remains open: a compatibility layer to bridge the semantic separation proved in 53 with the graph's provenance gap identified in 54A. First file to inspect when resuming: `lib/graph/schema.js` for current node fields, then `lib/graph/apply-proposal.js` for the production update path that populates nodes during inference turns.