{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2609.12658", "kind": "paper", "url": "https://arxiv.org/abs/2609.12658"}, {"id": "gh:v0yager33/proactivebench", "kind": "repo", "url": "https://github.com/v0yager33/ProactiveBench"}], "authors": ["Kaixuan Du", "Xin Wan", "Hang Zhang", "Meng Cao", "Dai Guan", "Ming Chen", "YuKun Wang"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2609.12658", "source_url": "https://arxiv.org/abs/2609.12658"}, "categories": ["General AI", "Vision & 3D", "cs.LG"], "description": {"en": "ProactiveBench evaluates models at one-second stream intervals without an explicit response cue in streaming video settings."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2609.12658", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2609.12658", "title": "ProactiveBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Alibaba Group (Cayman Islands)", "Alibaba Group (United States)", "Alibaba Group (China)", "Beihang University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2609.12658", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2609.12658", "source_url": "https://openalex.org/W7212711481"}, "key": "claire-radar:2609.12658", "modality": null, "name": "ProactiveBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "false", "origin_record_id": "bm_proactivebench_d1f3f8ef", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "71bc7c2e20ab8a7d90d7e6556214ac6902e71f2cc6e2df65cc226c8b688a283c", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-09-14T06:44:31.931424Z", "source_url": "https://arxiv.org/abs/2609.12658"}, "publisher": null, "released": "2026-09-11", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2609.12658", "source_url": "https://arxiv.org/abs/2609.12658"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2609-12658", "source": "claire_radar", "source_benchmark_id": "2609.12658", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/v0yager33/ProactiveBench", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2609.12658", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"checkedAt": "2026-10-04T15:31:48.030916Z", "lastSuccessfulDate": "2026-09-15", "state": "stale"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "stale"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Streaming video interaction is a topical area, but the record supplies no verified artifacts, scoring details, or public reuse path to support a stronger forecast.", "score": 38}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-14T10:07:09.772218Z", "inputHash": "4eb90638854210e78f8ecf37223401c9444921466f84663c1f7ee924edc29db1", "model": "deepseek-v4-flash"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "We introduce ProactiveBench, which evaluates models at one-second stream intervals without an explicit response cue.", "canonicalNameSource": "abstract", "decisionReason": "The GitHub link is unverified and no artifact excerpt is provided, so the evaluation and reuse path cannot be established.", "evaluationEvidence": {"reuse": "Code and data are available at https://github.com/v0yager33/ProactiveBench, providing a public path for other teams to run the benchmark.", "scoring": "Event Sensitivity combines response and silence rates; four window-based subtasks distinguish early, in-window, and missed responses; Duplicate Counting penalizes omissions and repetitions.", "task": "Evaluate streaming video models on proactive interaction: monitor a standing request and respond within an appropriate interval after a target event while otherwise remaining silent."}, "model": "deepseek-v4-pro", "reviewedAt": "2026-09-14T06:44:31.931424Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "ProactiveBench evaluates models at one-second stream intervals without an explicit response cue in streaming video settings.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "public artifact URL"], "snippet": "We introduce ProactiveBench, which evaluates models at one-second stream intervals without an explicit response cue."}, "familyId": "bmf_b1897bd0920e", "firstSeenAt": "2026-09-14", "heat": null, "id": "bm_proactivebench_d1f3f8ef", "indexedAt": "2026-09-14T00:39:56.575498Z", "industrySectors": [], "links": {"code": "https://github.com/v0yager33/ProactiveBench", "data": null, "paper": "https://arxiv.org/abs/2609.12658", "pdf": "https://arxiv.org/pdf/2609.12658", "project": null, "report": "https://arxiv.org/abs/2609.12658"}, "metrics": [], "motivation": "Streaming video understanding requires models to process continuous multimodal input while maintaining temporal context.", "name": "ProactiveBench", "oneLine": "ProactiveBench evaluates models at one-second stream intervals without an explicit response cue in streaming video settings.", "paperTitle": "ProactiveBench: Can Streaming Video Models Really Interact Like Humans?", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2609.12658", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-14T00:39:56.575498Z"}, "publications": [], "ranking": {"30d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": 30, "historyDate": "2026-09-05", "rank": null, "score": null, "windowDays": 30}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.0, "value": 0}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 182, "score": 23}, "method": "attention-ranking-v13", "rank": 182, "score": 23}, "90d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.0, "value": 0}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 446, "score": 23}, "method": "attention-ranking-v13", "rank": 446, "score": 23}}, "readiness": "Runnable", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-09-11", "source": {"authors": ["Kaixuan Du", "Xin Wan", "YuKun Wang", "Hang Zhang", "Meng Cao", "Dai Guan", "Ming Chen", "Ni Li"], "categories": ["cs.LG"], "id": "2609.12658", "title": "ProactiveBench: Can Streaming Video Models Really Interact Like Humans?", "type": "arxiv", "url": "https://arxiv.org/abs/2609.12658"}, "sourceUpdatedAt": "2026-09-11T10:05:42Z", "topics": ["cs.LG"], "venueAttempts": [], "whyItMatters": "Streaming video interaction requires evaluating timely, unprompted model responses rather than cued outputs."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
