Metraipolis › archive › dc_3708eeca1d08

CompressionBench: How can your coding agents compress LLMs? — arXiv:2611.06858

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.06858
License (as archived)
CC0-1.0
Archived
2026-11-05T18:17:12.788Z
SHA-256 of the archived text
feb0c43fdf3cd727b5059d82b12d9a940a52e0fa3190e7c25c39a4a95cb03ea4
Stable URL of this page
https://metraipolis.com/documento/dc_3708eeca1d08

How to cite

CompressionBench: How can your coding agents compress LLMs? — arXiv:2611.06858 — arXiv (en), https://arxiv.org/abs/2611.06858 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3708eeca1d08 (Metraipolis archive dc_3708eeca1d08, sha256 feb0c43fdf3cd727b5059d82b12d9a940a52e0fa3190e7c25c39a4a95cb03ea4)
BibTeX
@misc{dc3708eeca1d08,
  title = {CompressionBench: How can your coding agents compress LLMs? — arXiv:2611.06858},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.06858},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3708eeca1d08; sha256 feb0c43fdf3cd727b5059d82b12d9a940a52e0fa3190e7c25c39a4a95cb03ea4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.06858. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recursive self-improvement requires AI systems to improve the models and infrastructure used by subsequent systems. LLM compression offers a concrete efficiency objective: retain a pretrained model's behavior within a smaller model storage budget. We study whether coding agents can carry out this step autonomously. We formulate compression as a terminal task in which an agent receives a source checkpoint, storage limit, inference environment, and time budget, then constructs compressed weights and executable inference code. The agent controls its algorithm and experiments; perplexity feedback guides search, and independent verification checks the final model. Across six agents and six model--size settings, 29 of 36 submissions pass verification. Selecting the lowest-perplexity agent design per task yields an average 3.4% perplexity increase over BF16 and 77.33% downstream accuracy, compared with 79.56% for BF16. Recorded experiments reveal how agents achieve these outcomes and where improvement breaks down. Agents lower PPL in 21 of 24 revisions at nearly fixed storage; complete-model feedback reverses 17 of 79 prior local rankings; and generation and inference tests expose failures that likelihood alone misses. The results identify representation design, experimental judgment, and capability validation as central to agent-driven model compression, a measurable component of efficiency-oriented self-improvement.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.06858 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-05T18:12:49.019Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).