Research documents
Research log
Append-only, auditable record of every question, experiment, result, and decision.
State of the art
Phase 2 — technique taxonomy with epistemic status (established / contested / debunked-in-part).
Research gaps
Phase 3 — ≥20 falsifiable candidate directions, each with its smallest Mac-runnable falsifying experiment.
Candidate ranking
Phase 4 — 10-axis scoring and 3–5 selected prototype candidates.
Methodology
Phase 9 — the formal pipeline turning (checkpoint + corpora + budget) into confidence-labeled atlas entries.
not yet writtenNovelty check
Phase 11 — final novelty verification (assume not novel until shown otherwise).
not yet writtenBibliography
Every consulted source with URL and access date.
Phase 1 literature notes
4.1 observational probing
Phase 1 notes — §4.1 Observational probing
4.10 visualization atlas design
Phase 1 notes — §4.10 Visualization and atlas prior art
4.2 causal interventions
Phase 1 notes — §4.2 Causal / interventional methods
4.3 feature decomposition
Phase 1 notes — §4.3 Feature decomposition (SAEs, superposition, dictionary methods)
4.4 circuits and graphs
Phase 1 notes — §4.4 Circuits and computational graphs
4.5 weight space cartography
Phase 1 notes — §4.5 Weight-space cartography (no forward pass required)
4.6 knowledge localization editing
Phase 1 notes — §4.6 Knowledge localization and editing
4.7 cross model comparison
Phase 1 notes — §4.7 Cross-model and cross-scale comparison
4.8 statistics epistemology
Phase 1 notes — §4.8 Statistics and epistemology of interpretability
4.9 systems local mapping
Phase 1 notes — §4.9 Systems side — local mapping at scale on macOS
The project charter itself is public: read the full research charter.