Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel · HackerPapers discussions

Smarter and Cheaper at Once: Byte-Exact KV-Cache Grafting Turns a Frozen Small Model into a Verified-Knowledge Flywheel — HN submission

Mon, 20 Jul 2026 08:48:06 GMT

Making models smarter and cheaper at the same time

https://arxiv.org/abs/2607.14431

2 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Mon, 20 Jul 2026 08:45:05 GMT

https://arxiv.org/abs/2607.14431, s

Paper and discussion history. A mention is not an endorsement.

Fri, 17 Jul 2026 02:37:51 GMT

Show HN: KV-Cache Grafting – Boosting frozen 12B LLMs to 93.3% AIME accuracy

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.