{"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-2606-27981","title":"ToxiREX: A Dataset on Toxic REasoning in ConteXt","arxiv_id":"2606.27981","date":"2026-06-26","proceeding":null,"authors":["Stefan F. Schouten","Ilia Markov","Piek Vossen"],"abstract":"We introduce a new, contextual, multilingual dataset called ToxiREX: Toxic REasoning in ConteXt. The dataset consists of threads of Reddit comments and structured characterizations of what the comments imply, following a systematic toxic reasoning schema developed in a previous paper. Using the schema allows us to capture and explain implicit and context-dependent toxicity, while supporting mappings to existing toxicity taxonomies. The dataset includes comments in six languages (English, Arabic, Turkish, Spanish, German, and Dutch), collected from posts connected to specific major events (e.g. the 2023 Turkey earthquakes; the Russian invasion of Ukraine). We describe the context-preserving preprocessing of the threads. We create a training set of 125 thousand comments which is annotated by a commercially available LLM, and a test set of just under three thousand comments that is annotated by native speakers. We show that apparent disagreements in the test set annotations often reflect defensible alternative interpretations rather than noise. Finally, we provide baseline results by prompting and fine-tuning language models. To produce these results, we develop evaluation strategies for our hierarchical, schema-based predictions. While models perform better than random, there remains a lot of room for improvement, showing the task to be challenging. ToxiREX is the first dataset to simultaneously incorporate multiple languages, conversational context, and implicit toxicity, while using the toxic reasoning schema for rich, structured annotations. Dataset available at: https://github.com/cltl/toxirex","url_abs":"https://arxiv.org/abs/2606.27981","url_pdf":"https://arxiv.org/pdf/2606.27981","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":{"syntology_url":"https://syntology.ai/paper/2606.27981","atlas_url":"https://app.syntology.ai/?focus=2606.27981","mcp":null,"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/cltl/toxirex","reach":{"status":"ok"}},{"mentioned_in_github":null,"is_official":null,"provenance":"deterministic:regex_extraction","mentioned_in_paper":null,"url":"https://github.com/FlorianPusse/Lingoturk","reach":{"status":"ok"}}],"summary":{},"by_repo_kind":{},"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":[]},"arxiv_metadata":{"licence":"arXiv metadata, CC0 1.0 (https://info.arxiv.org/help/license)","fields":["title","abstract","authors","date"],"primary_category":"cs.CL","source":"arxiv_2026.jsonl"},"syntology_extracted_results":{"kind":"leaderboard_placements","source":"Syntology's leaderboard-shaped extractor over the paper's own arXiv-HTML tables: a model pointed at a cell, the number was read from that cell and checked against the board's metric, dataset, split and scale, and an independent check accepted the entry; not reviewed by the paper's authors or the archive's editors","extractor_model":"global.anthropic.claude-sonnet-4-5-20250929-v1:0","verifier_model":null,"prompt_sha":"fa63d4bb9d755694","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","papers_newer_than_archive":9623,"papers_checked":6885},"entries":[],"not_placed":{"boards":0,"rejected_by_independent_check":0,"refused_by_a_rule":0,"check_did_not_answer":0,"proposed_without_a_cell":0,"declined_by_site":0}}}