{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.02642", "kind": "paper", "url": "https://arxiv.org/abs/2606.02642"}, {"kind": "website", "url": "https://chenshuang-zhang.github.io/projects/svhalluc/"}], "authors": ["Chenshuang Zhang", "Kyeong Seon Kim", "Chengxin Liu", "Tae-Hyun Oh"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.02642", "source_url": "https://arxiv.org/abs/2606.02642"}, "categories": ["General AI", "Multimodal", "eess.AS"], "description": {"en": "SVHalluc evaluates speech-vision hallucination in audio-visual large language models, focusing on semantic and temporal alignment between speech content and visual signals."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.02642", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.02642", "title": "SVHalluc"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Korea Advanced Institute of Science and Technology"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.02642", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.02642", "source_url": "https://openalex.org/W7163333871"}, "key": "claire-radar:2606.02642", "modality": null, "name": "SVHalluc", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_svhalluc_d54e2d9e", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "a7d7acabac9f3591a7de0ca700aed08fa591d4b2f47a072ddb8957e469d75fd3", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.02642"}, "publisher": null, "released": "2026-05-31", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.02642", "source_url": "https://arxiv.org/abs/2606.02642"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-02642", "source": "claire_radar", "source_benchmark_id": "2606.02642", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Multimodal", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2606.02642", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-30", "state": "stale"}}, "status": "stale"}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "f1bed85e3bfe08a642a1a3922c8487c37af6d1c0eacd88b9a6a4d7002c669bed", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "SVHalluc evaluates speech-vision hallucination in audio-visual large language models, focusing on semantic and temporal alignment between speech content and visual signals.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "To systematically study this, we introduce SVHalluc, the first comprehensive benchmark for evaluating speech-vision hallucination in audio-visual LLMs."}, "familyId": "bmf_d73ceabecf48", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 1, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "candidates"}, "heat": null, "id": "bm_svhalluc_d54e2d9e", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.02642", "paper": "https://arxiv.org/abs/2606.02642", "pdf": "https://arxiv.org/pdf/2606.02642", "project": "https://chenshuang-zhang.github.io/projects/svhalluc/", "report": "https://arxiv.org/abs/2606.02642"}, "metrics": [], "motivation": "Despite the success of audio-visual large-language models (LLMs), they can produce plausible but ungrounded outputs, termed hallucination.", "name": "SVHalluc", "oneLine": "SVHalluc evaluates speech-vision hallucination in audio-visual large language models, focusing on semantic and temporal alignment between speech content and visual signals.", "paperTitle": "SVHalluc: Benchmarking Speech-Vision Hallucination in Audio-Visual Large Language Models", "primaryDomain": "General AI", "publication": {"evidence": "Accepted at CVPR 2026", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2606.02642", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "CVPR 2026", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "Chenshuang Zhang et al.", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://chenshuang-zhang.github.io/projects/svhalluc/"}], "ranking": {}, "readiness": "Inspectable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-31", "source": {"authors": ["Chenshuang Zhang", "Kyeong Seon Kim", "Chengxin Liu", "Tae-Hyun Oh"], "categories": ["eess.AS", "cs.AI", "cs.CV", "cs.LG", "cs.MM", "cs.SD"], "id": "2606.02642", "title": "SVHalluc: Benchmarking Speech-Vision Hallucination in Audio-Visual Large Language Models", "type": "arxiv", "url": "https://arxiv.org/abs/2606.02642"}, "sourceUpdatedAt": "2026-05-31T14:00:37Z", "topics": ["eess.AS"], "venueAttempts": [{"decisionRaw": "Accepted at CVPR 2026", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "Accepted at CVPR 2026", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2606.02642"}], "reviewStatus": "accepted", "venueName": "CVPR 2026"}], "whyItMatters": "Addresses a critical gap in evaluating audio-visual LLMs, as prior benchmarks ignored speech-induced hallucinations. Provides a systematic method to assess model grounding, aiding in model development and selection for multimodal applications."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
