{"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-2604-20276","title":"Rethinking Intrinsic Dimension Estimation in Neural Representations","arxiv_id":"2604.20276","date":"2026-04-22","proceeding":null,"authors":["Rickmer Schulte","David Rügamer"],"abstract":"The analysis of neural representation has become an integral part of research aiming to better understand the inner workings of neural networks. While there are many different approaches to investigate neural representations, an important line of research has focused on doing so through the lens of intrinsic dimensions (IDs). Although this perspective has provided valuable insights and stimulated substantial follow-up research, important limitations of this approach have remained largely unaddressed. In this paper, we highlight a crucial discrepancy between theory and practice of IDs in neural representations, theoretically and empirically showing that common ID estimators are, in fact, not tracking the true underlying ID of the representation. We contrast this negative result with an investigation of the underlying factors that may drive commonly reported ID-related results on neural representation in the literature. Building on these insights, we offer a new perspective on ID estimation in neural representations.","url_abs":"https://arxiv.org/abs/2604.20276","url_pdf":"https://arxiv.org/pdf/2604.20276","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=2604.20276","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2604.20276"}},"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/rickmer-schulte/rethinking-neural-id","reach":null}],"summary":{"unverified":14},"by_repo_kind":{"found_in_text":{"samples":14,"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":"3ce2825a9b946af8","entry":"as_float_array","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/plotting/llm.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/plotting/llm.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":"3ce2825a9b946af8"}},{"code_sha256_prefix":"17f3b16e5ea15dc2","entry":"ensure_parent","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/artifacts.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/artifacts.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":"17f3b16e5ea15dc2"}},{"code_sha256_prefix":"e49a00e7ca03f0b2","entry":"entropy","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/layerwise.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/layerwise.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":"e49a00e7ca03f0b2"}},{"code_sha256_prefix":"576ef035a54078ea","entry":"est_entropy","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/cnn/entropy.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/cnn/entropy.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":"576ef035a54078ea"}},{"code_sha256_prefix":"3e16c6d39b7e49bf","entry":"get_image_dataset_dir","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/registry.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/registry.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":"3e16c6d39b7e49bf"}},{"code_sha256_prefix":"dc4bfe9e3be75fb2","entry":"get_llm_model_spec","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/registry.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/registry.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":"dc4bfe9e3be75fb2"}},{"code_sha256_prefix":"73d628f000e0b0b2","entry":"get_vit_model_spec","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/registry.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/registry.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":"73d628f000e0b0b2"}},{"code_sha256_prefix":"94afb00cd15efed7","entry":"knn_avg_distances","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/layerwise.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/layerwise.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":"94afb00cd15efed7"}},{"code_sha256_prefix":"ff7dd6d17d609bc6","entry":"load_representation_dict","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/artifacts.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/artifacts.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":"ff7dd6d17d609bc6"}},{"code_sha256_prefix":"d5c011367c4c2057","entry":"mean_var_pairwise_cosine","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/layerwise.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/layerwise.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":"d5c011367c4c2057"}},{"code_sha256_prefix":"6f8ecb5324ab7242","entry":"normalize_gram","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/cnn/entropy.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/cnn/entropy.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":"6f8ecb5324ab7242"}},{"code_sha256_prefix":"d2cb04fa1362a926","entry":"relative_depth","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/plotting/vit.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/plotting/vit.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":"d2cb04fa1362a926"}},{"code_sha256_prefix":"73da6236655d107e","entry":"save_representation_dict","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/artifacts.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/artifacts.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":"73da6236655d107e"}},{"code_sha256_prefix":"5dd219da23b99f91","entry":"von_neumann_entropy","repo":"rickmer-schulte/rethinking-neural-id","repo_kind":"found_in_text","path":"src/rethinking_neural_id/metrics/cnn/entropy.py","file_url":"https://github.com/rickmer-schulte/rethinking-neural-id/blob/HEAD/src/rethinking_neural_id/metrics/cnn/entropy.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":"5dd219da23b99f91"}}]},"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}