{"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-2604-04287","title":"Entropy, Disagreement, and the Limits of Foundation Models in Genomics","arxiv_id":"2604.04287","date":"2026-04-05","proceeding":null,"authors":["Maxime Rochkoulets","Lovro Vrček","Mile Šikić"],"abstract":"Foundation models in genomics have shown mixed success compared to their counterparts in natural language processing. Yet, the reasons for their limited effectiveness remain poorly understood. In this work, we investigate the role of entropy as a fundamental factor limiting the capacities of such models to learn from their training data and develop foundational capabilities. We train ensembles of models on text and DNA sequences and analyze their predictions, static embeddings, and empirical Fisher information flow. We show that the high entropy of genomic sequences -- from the point of view of unseen token prediction -- leads to near-uniform output distributions, disagreement across models, and unstable static embeddings, even for models that are matched in architecture, training and data. We then demonstrate that models trained on DNA concentrate Fisher information in embedding layers, seemingly failing to exploit inter-token relationships. Our results suggest that self-supervised training from sequences alone may not be applicable to genomic data, calling into question the assumptions underlying current methodologies for training genomic foundation models.","url_abs":"https://arxiv.org/abs/2604.04287","url_pdf":"https://arxiv.org/pdf/2604.04287","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=2604.04287","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2604.04287"}},"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/lbcb-sci/GFMs","reach":null}],"summary":{"unverified":11},"by_repo_kind":{"found_in_text":{"samples":11,"ran":0,"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":"20395d9e9c5580b0","entry":"barplot","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/analyze/plotting.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/analyze/plotting.py","link_basis":"first_harvest_node","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":"20395d9e9c5580b0"}},{"code_sha256_prefix":"b24143af5974d16b","entry":"choose_longest_seqs","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/train/data.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/train/data.py","link_basis":"first_harvest_node","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":"b24143af5974d16b"}},{"code_sha256_prefix":"83f5b11561f82af9","entry":"collect","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"discriminator/plot_logits.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/discriminator/plot_logits.py","link_basis":"first_harvest_node","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":"83f5b11561f82af9"}},{"code_sha256_prefix":"e84398d63ab4ce8b","entry":"compute_weights","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"discriminator/inference.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/discriminator/inference.py","link_basis":"first_harvest_node","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":"e84398d63ab4ce8b"}},{"code_sha256_prefix":"3246e96fb267885d","entry":"cut","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"discriminator/inference.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/discriminator/inference.py","link_basis":"first_harvest_node","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":"3246e96fb267885d"}},{"code_sha256_prefix":"b39fafeed31cde18","entry":"effective_sample_size","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"discriminator/plot_logits.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/discriminator/plot_logits.py","link_basis":"first_harvest_node","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":"b39fafeed31cde18"}},{"code_sha256_prefix":"f22063481f5a3d06","entry":"jaccard","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/analyze/metrics.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/analyze/metrics.py","link_basis":"first_harvest_node","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":"f22063481f5a3d06"}},{"code_sha256_prefix":"1d8286886d68bf4d","entry":"order_fisher","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/analyze/plotting.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/analyze/plotting.py","link_basis":"first_harvest_node","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":"1d8286886d68bf4d"}},{"code_sha256_prefix":"88f683469afe4ec1","entry":"order_fisher_full","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/analyze/plotting.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/analyze/plotting.py","link_basis":"first_harvest_node","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":"88f683469afe4ec1"}},{"code_sha256_prefix":"2a2b18825f81ed07","entry":"sample_is_clean","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"discriminator/preprocess.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/discriminator/preprocess.py","link_basis":"first_harvest_node","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":"2a2b18825f81ed07"}},{"code_sha256_prefix":"ef9c89e18e06f4d2","entry":"topk_overlap","repo":"lbcb-sci/GFMs","repo_kind":"found_in_text","path":"src/analyze/metrics.py","file_url":"https://github.com/lbcb-sci/GFMs/blob/HEAD/src/analyze/metrics.py","link_basis":"first_harvest_node","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":"ef9c89e18e06f4d2"}}]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.LG","source":"arxiv_2026.jsonl"},"syntology_extracted_results":null}