of New Haven · small models, real problems.
We are a small local-model research program studying how much reasoning, composition, and adaptation can be obtained under practical compute limits — and how to carry earned improvements into an offline-capable assistant.
NHV-MARK-REV-A
compression glyph · signal unit
Under bounded training and inference resources, which allocations of computation, representation, memory, and reusable skills improve transfer to unfamiliar tasks?
The R03 evaluation apparatus is merged, the R02 readiness path has passed its reviewed smoke, and the E01 records baseline protocol and runner are frozen. The first controlled E01 result is still pending; no experiment score has been published. We count failures, resource limits, and contradictions as part of the record.
The first controlled baseline is frozen: 32 deterministic structured-record episodes, canonical records plus explicit primitive operations, and an independent exact-match scorer. Execution is the next gate; no result is claimed yet.
Read E01 protocol ↗Can reusing parameters improve reasoning beyond ordinary depth or additional answer sampling? The initial work separates mechanism evidence from practical deployment claims.
Read E02 protocol ↗When a task introduces a new rule, when does a bounded update help more than examples in context, retrieval, or explicit rule search — and what does reset cost?
Read E03 protocol ↗A score is not a theory. A pleasant demo is not a transfer result. The laboratory publishes the protocol, provenance, costs, errors, and limits alongside any claimed gain.
“Keep development feedback separate from final evaluation. Preserve unsuccessful runs and contradictions. A small clean negative result is a legitimate deliverable.”
Research conduct / current policyWe count model weights, caches, retrieval, learned libraries, verifiers, preprocessing, CPU memory, wall time, and external services. “Small” is a system property.
Issues, branches, reviews, and decision records live in the canonical Forgejo forge. The GitHub repository is the public mirror for readers and discoverability. If you find a mistake in a claim or a protocol, that is useful research feedback.