Metraipolis › archive › dc_cf976a60af91
CogAdapt: Cognition-informed Sparse Adaptation of Code LLMs — arXiv:2610.07446
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.07446
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T11:43:33.747Z
- SHA-256 of the archived text
e17897694f22484470f312d94f58c68516295a827eb47541c7adecf55318eb79- Stable URL of this page
https://metraipolis.com/documento/dc_cf976a60af91
How to cite
CogAdapt: Cognition-informed Sparse Adaptation of Code LLMs — arXiv:2610.07446 — arXiv (en), https://arxiv.org/abs/2610.07446 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_cf976a60af91 (Metraipolis archive dc_cf976a60af91, sha256 e17897694f22484470f312d94f58c68516295a827eb47541c7adecf55318eb79)
BibTeX
@misc{dccf976a60af91,
title = {CogAdapt: Cognition-informed Sparse Adaptation of Code LLMs — arXiv:2610.07446},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.07446},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_cf976a60af91; sha256 e17897694f22484470f312d94f58c68516295a827eb47541c7adecf55318eb79},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.07446. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large language models (LLMs) have become increasingly capable of generating code. However, achieving stronger code-generation performance still often relies on costly model adaptation, i.e., fine-tuning pretrained model parameters. Prior studies have shown correspondence between human code processing and neural models' attention or internal computation. Human-aligned learning approaches use cognitive signals to guide training, but typically adapt a large portion of the model, leaving training costs largely unchanged. Human cognitive signals may indicate not only what the model must learn from, but also where adaptation is most useful. We investigate whether human responses during code reading correspond to code-model behavior and can guide selective adaptation without sacrificing performance. We present CogAdapt, a cognition-informed framework for task-dependent sparse adaptation of code models. CogAdapt first learns transferable program-level and token-level priors from human Electroencephalography (EEG) and attention data, then combines these priors with the frozen model's response to each coding task to determine how much adaptation to allocate and which transformer blocks should receive updates. During fine-tuning, only the selected blocks are updated, while no new human recordings are required for inference. Across Qwen and GLM, we find consistent correspondence between human reading behavior and Mixture-of-Experts (MoE) computation. CogAdapt achieves the best pass@1 across both LiveCodeBench and BigCodeBench, including gains of 10.86 and 6.29 percentage points over matched regular fine-tuning on LiveCodeBench, while reducing gradient-eligible adaptation parameters by 86.21-87.21%. These results suggest that human comprehension signals can provide useful guidance for making code-model adaptation both more selective and more effective. Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.SE, nuovi lavori del giorno). Categorie arXiv: Software Engineering (cs.SE). Pagina del lavoro: https://arxiv.org/abs/2610.07446 — 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 2026-10-07T11:19:39.502Z dalla pagina https://arxiv.org/list/cs.SE/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).