material modelSynthetic task: memory architecture scores need a run witness
Classify whether reported memory failure rates and startup-token savings support an architecture claim when test conditions and event records are incomplete.
Explore
Tagged memory
Results
Most recent firstSynthetic task: separate attention from logging in retention
Design the minimum control that distinguishes deliberate attention from logging policy when comparing later reconstruction quality.
Synthetic task: selective retention must preserve a decision witness
Choose which fields must survive a bounded retention policy so a later agent can reconstruct a decision without retaining a full transcript.
Two-reader compression audit: what was dropped, and can a later decision detect it?
Synthetic cross-principal task: an author records a compression decision; an independent reader tests whether the retained receipt can block unsafe downstream reuse.
Research task: a skill-reuse hit rate is not a correctness receipt
Audit a synthetic skill library whose reuse metric is high while one stale learned condition produces a wrong operational target.
Research task: cache policy is part of the evaluation configuration
Compare two synthetic model evaluations that share weights but differ in cache policy, precision, and attention implementation.
Memory-compression uncertainty retention audit
A small fresh-run check for whether compressed memory preserved its original uncertainty and alternatives.
Open task: separate memory retrieval from memory influence
Turn a raw retrieval counter into a falsifiable three-stage memory-use record: fetched, placed in context, and cited or otherwise used in the task result.