{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/arxiv-2601-08517","title":"Closed-Loop LLM Discovery of Non-Standard Channel Priors in Vision Models","arxiv_id":"2601.08517","date":"2026-01-13","proceeding":null,"authors":["Tolgay Atinc Uzun","Dmitry Ignatov","Radu Timofte"],"abstract":"Channel-configuration search, the optimization of layer specifications such as channel widths in deep neural networks, presents a combinatorial challenge constrained by tensor-shape compatibility and computational budgets. We investigate whether large language models (LLMs) can support neural architecture search (NAS) by reasoning over architectural code structures in ways that complement traditional search heuristics. We apply an LLM-driven NAS framework to channel-configuration search, formulating the task as conditional code generation in which the LLM refines architectural specifications using performance feedback. To address data scarcity, we generate a corpus of valid, shape-consistent architectures through abstract syntax tree (AST) mutations. Although these mutated networks are not necessarily optimized for performance, they provide structural examples that help the LLM learn executable architectural patterns and relate channel configurations to model performance. Experimental results on CIFAR-100 show that the closed-loop LLM improves upon the initial AST-generated architecture population under the same proxy-evaluation protocol. Our analysis further shows that the generated architectures reflect domain-specific design patterns, including non-standard channel widths and late-stage expansion, highlighting the potential of language-driven design for code-level NAS. The code and prompts are publicly available at https://github.com/ABrain-One/NN-GPT, and the generated deep neural networks are published at https://github.com/ABrain-One/NN-Dataset under model names with the prefix ast-dimension-.","url_abs":"https://arxiv.org/abs/2601.08517","url_pdf":"https://arxiv.org/pdf/2601.08517","source":{"archive":null,"snapshot":"2025-07-28","note":"not in the Papers with Code archive (frozen at the snapshot)","row_kind":"graph","title_abstract_authors_date":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)"},"code_links":[],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2601.08517","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2601.08517"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/ABrain-One/NN-Dataset","reach":null},{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/ABrain-One/NN-GPT","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":4,"unverified":3},"by_repo_kind":{"found_in_text":{"samples":7,"ran":4,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"0c65975ee3f717f6","entry":"build_hp_dict","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/brute/llr/layerwise_lr.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/brute/llr/layerwise_lr.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0c65975ee3f717f6"}},{"code_sha256_prefix":"a3b3192b3772f883","entry":"extract_accuracy","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/act/edge/EdgeScore.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/act/edge/EdgeScore.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a3b3192b3772f883"}},{"code_sha256_prefix":"a9e7d837fc20beb9","entry":"get_supported_hp_from_source","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/brute/llr/layerwise_lr.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/brute/llr/layerwise_lr.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a9e7d837fc20beb9"}},{"code_sha256_prefix":"ab8fb944152a9e91","entry":"infer_arch_family","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/act/AccPredictor.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/act/AccPredictor.py","link_basis":"plan_row","language":"python","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"ab8fb944152a9e91"}},{"code_sha256_prefix":"c292f613d08d318a","entry":"code_minhash","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/act/edge/EdgeScore.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/act/edge/EdgeScore.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c292f613d08d318a"}},{"code_sha256_prefix":"c835268f24f02d57","entry":"data_preprocessing","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/act/AccPredictor.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/act/AccPredictor.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"c835268f24f02d57"}},{"code_sha256_prefix":"1dbabd335fc4dd31","entry":"max_similarity","repo":"ABrain-One/NN-GPT","repo_kind":"found_in_text","path":"ab/gpt/act/edge/EdgeScore.py","file_url":"https://github.com/ABrain-One/NN-GPT/blob/HEAD/ab/gpt/act/edge/EdgeScore.py","link_basis":"plan_row","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1dbabd335fc4dd31"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CV","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}