{"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/dialectal-toxicity-detection-evaluating-llm","title":"Dialectal Toxicity Detection: Evaluating LLM-as-a-Judge Consistency Across Language Varieties","arxiv_id":"2411.10954","date":"2024-11-17","proceeding":null,"authors":["Fahim Faisal","Md Mushfiqur Rahman","Antonios Anastasopoulos"],"abstract":"There has been little systematic study on how dialectal differences affect toxicity detection by modern LLMs. Furthermore, although using LLMs as evaluators (\"LLM-as-a-judge\") is a growing research area, their sensitivity to dialectal nuances is still underexplored and requires more focused attention. In this paper, we address these gaps through a comprehensive toxicity evaluation of LLMs across diverse dialects. We create a multi-dialect dataset through synthetic transformations and human-assisted translations, covering 10 language clusters and 60 varieties. We then evaluated three LLMs on their ability to assess toxicity across multilingual, dialectal, and LLM-human consistency. Our findings show that LLMs are sensitive in handling both multilingual and dialectal variations. However, if we have to rank the consistency, the weakest area is LLM-human agreement, followed by dialectal consistency. Code repository: \\url{https://github.com/ffaisal93/dialect_toxicity_llm_judge}","url_abs":"https://arxiv.org/abs/2411.10954v1","url_pdf":"https://arxiv.org/pdf/2411.10954v1.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":"dialectal-toxicity-detection-evaluating-llm","repo_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2411.10954","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2411.10954"}},"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/ffaisal93/dialect_toxicity_llm_judge","reach":{"status":"ok"}}],"summary":{"unverified":16},"by_repo_kind":{"official":{"samples":16,"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":16,"samples":[{"code_sha256_prefix":"62688b75832990a2","entry":"add_avg_row_and_column","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/analyze_toxicity_shift.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/analyze_toxicity_shift.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"62688b75832990a2"}},{"code_sha256_prefix":"b08cdf16d602f1bd","entry":"analyze_toxicity_shift_cluster","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/analyze_toxicity_shift.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/analyze_toxicity_shift.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b08cdf16d602f1bd"}},{"code_sha256_prefix":"9d587618cff0b7b2","entry":"assign_bins","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/compute_consistency.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/compute_consistency.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"9d587618cff0b7b2"}},{"code_sha256_prefix":"0c735f574ddfa21c","entry":"build_combined_toxicity_table","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/analyze_toxicity_shift.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/analyze_toxicity_shift.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"0c735f574ddfa21c"}},{"code_sha256_prefix":"3c892580bbb50de4","entry":"clean_discrepancies","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/compute_consistency.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/compute_consistency.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"3c892580bbb50de4"}},{"code_sha256_prefix":"53a67568c34824df","entry":"compute_bins_and_metrics","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/compute_score.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/compute_score.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"53a67568c34824df"}},{"code_sha256_prefix":"4ace4e4a4187e874","entry":"compute_metrics","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/compute_consistency.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/compute_consistency.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"4ace4e4a4187e874"}},{"code_sha256_prefix":"7db0225118cee3f5","entry":"format_output","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/llm_support.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/llm_support.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7db0225118cee3f5"}},{"code_sha256_prefix":"86a53fbc72e76e64","entry":"formulate_chat_dict","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/prompt_support.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/prompt_support.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"86a53fbc72e76e64"}},{"code_sha256_prefix":"b5e5fa88cd1c0091","entry":"formulate_prompt","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/prompt_support.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/prompt_support.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b5e5fa88cd1c0091"}},{"code_sha256_prefix":"e83911a29d37549b","entry":"get_output_format","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/prompt_support.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/prompt_support.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"e83911a29d37549b"}},{"code_sha256_prefix":"7c243811cbb0d6b1","entry":"get_score_distribution","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/annotator_toxicity.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/annotator_toxicity.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"7c243811cbb0d6b1"}},{"code_sha256_prefix":"39bd94b7698cf5a8","entry":"get_token","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/utils.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"39bd94b7698cf5a8"}},{"code_sha256_prefix":"62ed0146e29e7bb7","entry":"load_config","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/utils.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"62ed0146e29e7bb7"}},{"code_sha256_prefix":"a0c7393409fd0be9","entry":"load_json","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/judge_translation_vllm.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/judge_translation_vllm.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"a0c7393409fd0be9"}},{"code_sha256_prefix":"b98daea8dee675ee","entry":"update_arguments_with_config","repo":"ffaisal93/dialect_toxicity_llm_judge","repo_kind":"official","path":"src/llm_judge/utils.py","file_url":"https://github.com/ffaisal93/dialect_toxicity_llm_judge/blob/HEAD/src/llm_judge/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"mcp_get_code":{"code_sha256":"b98daea8dee675ee"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}