{"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/ravel-evaluating-interpretability-methods-on","title":"RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations","arxiv_id":"2402.17700","date":"2024-02-27","proceeding":null,"authors":["Jing Huang","Zhengxuan Wu","Christopher Potts","Mor Geva","Atticus Geiger"],"abstract":"Individual neurons participate in the representation of multiple high-level concepts. To what extent can different interpretability methods successfully disentangle these roles? To help address this question, we introduce RAVEL (Resolving Attribute-Value Entanglements in Language Models), a dataset that enables tightly controlled, quantitative comparisons between a variety of existing interpretability methods. We use the resulting conceptual framework to define the new method of Multi-task Distributed Alignment Search (MDAS), which allows us to find distributed representations satisfying multiple causal criteria. With Llama2-7B as the target language model, MDAS achieves state-of-the-art results on RAVEL, demonstrating the importance of going beyond neuron-level analyses to identify features distributed across activations. We release our benchmark at https://github.com/explanare/ravel.","url_abs":"https://arxiv.org/abs/2402.17700v2","url_pdf":"https://arxiv.org/pdf/2402.17700v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"ravel-evaluating-interpretability-methods-on","repo_url":"https://github.com/explanare/ravel","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"attribute","task_name":"Attribute"},{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2402.17700","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2402.17700"}},"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":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/explanare/ravel","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"ran":12,"unverified":5},"by_repo_kind":{"official":{"samples":17,"ran":12,"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":"39d29ba2dce082a9","entry":"compute_cross_entropy_loss","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/metric_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/metric_utils.py","link_basis":"first_harvest_node","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":"39d29ba2dce082a9"}},{"code_sha256_prefix":"618cb3bd9cfb4a14","entry":"compute_disentangle_score","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/metric_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/metric_utils.py","link_basis":"first_harvest_node","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":"618cb3bd9cfb4a14"}},{"code_sha256_prefix":"5500053238475cfe","entry":"compute_metrics","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/metric_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/metric_utils.py","link_basis":"first_harvest_node","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":"5500053238475cfe"}},{"code_sha256_prefix":"ebcb0699723a970c","entry":"compute_principal_component","repo":"explanare/ravel","repo_kind":"official","path":"src/methods/pca.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/methods/pca.py","link_basis":"first_harvest_node","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":"ebcb0699723a970c"}},{"code_sha256_prefix":"e7618272c2024815","entry":"generate_batched","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/generation_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/generation_utils.py","link_basis":"harvester_set","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":"e7618272c2024815"}},{"code_sha256_prefix":"4637704289ff6a8a","entry":"get_score","repo":"explanare/ravel","repo_kind":"official","path":"src/methods/linear_adversarial_probe.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/methods/linear_adversarial_probe.py","link_basis":"first_harvest_node","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":"4637704289ff6a8a"}},{"code_sha256_prefix":"9a0c2776f5ac2bc0","entry":"init_sgd_classifier","repo":"explanare/ravel","repo_kind":"official","path":"src/methods/linear_adversarial_probe.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/methods/linear_adversarial_probe.py","link_basis":"first_harvest_node","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":"9a0c2776f5ac2bc0"}},{"code_sha256_prefix":"f648fddc170ec6d8","entry":"kept_first_n_label_token","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/dataset_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/dataset_utils.py","link_basis":"first_harvest_node","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":"f648fddc170ec6d8"}},{"code_sha256_prefix":"f404f9aad2c1b27e","entry":"kept_first_n_label_token_multitask","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/dataset_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/dataset_utils.py","link_basis":"first_harvest_node","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":"f404f9aad2c1b27e"}},{"code_sha256_prefix":"39872fcdf40c8524","entry":"preproc_tokenize","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/dataset_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/dataset_utils.py","link_basis":"first_harvest_node","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":"39872fcdf40c8524"}},{"code_sha256_prefix":"fe8f9421a66f408f","entry":"remove_invalid_token_id","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/intervention_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/intervention_utils.py","link_basis":"first_harvest_node","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":"fe8f9421a66f408f"}},{"code_sha256_prefix":"be76384e0b33672a","entry":"select_features_with_classifier","repo":"explanare/ravel","repo_kind":"official","path":"src/methods/select_features.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/methods/select_features.py","link_basis":"first_harvest_node","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":"be76384e0b33672a"}},{"code_sha256_prefix":"8240c646692f8698","entry":"gen_context_test_split","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/generate_ravel_instance.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/generate_ravel_instance.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":"8240c646692f8698"}},{"code_sha256_prefix":"58779741197dded2","entry":"gen_entity_test_split","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/generate_ravel_instance.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/generate_ravel_instance.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":"58779741197dded2"}},{"code_sha256_prefix":"8289a1e4195d682d","entry":"gen_train_split","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/generate_ravel_instance.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/generate_ravel_instance.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":"8289a1e4195d682d"}},{"code_sha256_prefix":"5eb6f7201a4d2f1c","entry":"symmetric","repo":"explanare/ravel","repo_kind":"official","path":"src/methods/linear_adversarial_probe.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/methods/linear_adversarial_probe.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":"5eb6f7201a4d2f1c"}},{"code_sha256_prefix":"01750c9c0ff161d4","entry":"train_intervention_step","repo":"explanare/ravel","repo_kind":"official","path":"src/utils/intervention_utils.py","file_url":"https://github.com/explanare/ravel/blob/HEAD/src/utils/intervention_utils.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":"01750c9c0ff161d4"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}