industry case study

Bonsai 2 27B

Ternary representation, packing overhead, activation transforms, kernel support, artifact size, runtime memory, and capability retention can rank a model differently depending on the deployment constraint. File size is not peak runtime memory, and retention relative to a source model does not establish frontier parity.

Disposition
study only
Study class
industry case study
Reviewed
2026-09-23
Evidence
Tracked source record

What mattered

Ternary representation, packing overhead, activation transforms, kernel support, artifact size, runtime memory, and capability retention can rank a model differently depending on the deployment constraint.

What changed locally

The retained exercise is a same-Mac comparison sheet covering revisions, task gates, peak RAM, time to first token, prefill, decode throughput, and total task latency.

Evidence limits

File size is not peak runtime memory, and retention relative to a source model does not establish frontier parity.

Retained disposition

A future comparison requires owner selection, bounded resources, and exact runtime and weight revisions.

The action is retained as study context rather than an active backlog item. Any implementation, download, training run, or benchmark requires a fresh scoped question under the repository's experiment gate.

How to read this dossier

This canonical page separates the retained repository decision from the material that informed it. A source may describe an external claim, historical measurement, local observation, or planned procedure; those evidence classes are not interchangeable. Follow the provenance links before reusing a number or method.

Status and disposition describe what PostTrainLLM retained when this record was indexed. They are not a live product promise, a newly run benchmark, or permission to restart historical work. Unknown or unmeasured fields remain unknown rather than being treated as zero.

To reuse the record, first name the exact claim you need and trace it to the linked source. Then check whether the original environment, model revision, data split, evaluator, hardware, and budget match the proposed use. If they do not, treat the record as a hypothesis or design reference and run the smallest fresh comparison that can falsify it. Preserve negative outcomes and regressions beside any improvement; a local win on one slice does not silently become a general capability claim.

Source provenance

The normalized record comes from docs/studies/registry.json. The links below are the tracked evidence and explanatory sources preserved with the record.