{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2609.12439", "kind": "paper", "url": "https://arxiv.org/abs/2609.12439"}], "authors": ["Liang Zhao", "Yong Wang", "Jiangzhe Chen"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2609.12439", "source_url": "https://arxiv.org/abs/2609.12439"}, "categories": ["Agents", "General AI", "Language & Knowledge"], "description": {"en": "A diagnostic benchmark for auditing citation-like artifacts in retrieval-augmented generation and agent-workflow evaluation, including content-equivalent pairs, citation ablations, correctness conflicts, quality gaps, prompt-strength ladders, decoupled judging, and workflow ranking."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2609.12439", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2609.12439", "title": "TraceJudgeBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2609.12439", "modality": null, "name": "TraceJudgeBench", "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_tracejudgebench_6d9b999a", "origin_source": "arxiv", "recognition_confidence": "0.85", "record_sha256": "d3597634659fb9d1f4278773e5fa9c9416a7a6df679128578fb3cf92d55bb81b", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-09-19T08:55:40.500135Z", "source_url": "https://arxiv.org/abs/2609.12439"}, "publisher": null, "released": "2026-09-11", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2609.12439", "source_url": "https://arxiv.org/abs/2609.12439"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2609-12439", "source": "claire_radar", "source_benchmark_id": "2609.12439", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "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": null, "hfPaperUrl": "https://huggingface.co/papers/2609.12439", "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", "state": "unavailable"}}, "status": "unavailable"}, "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "Limited artifact evidence and a niche diagnostic benchmarking topic suggest modest initial attention.", "score": 15}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-19T08:55:40.500135Z", "inputHash": "de63c1dc4cb4a1cbc61bb36ded593b71633332c45e3446bd62c412dcb41fc0c7", "model": "deepseek-v4-pro"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "We introduce TraceJudgeBench, a diagnostic benchmark for auditing citation-like artifacts in RAG and agent-workflow evaluation, covering content-equivalent pairs, citation ablations, correctness conflicts, human-validated soft and moderate quality gaps, prompt-strength ladders, decoupled judging, and a controlled workflow-ranking probe.", "canonicalNameSource": "abstract", "decisionReason": "Only the abstract is provided; there is no evidence of released task inputs, scoring rules, or a public reuse path required for admission.", "evaluationEvidence": {"reuse": "The supplementary artifact contains benchmark splits, prompts, raw judge outputs, validation summaries, and analysis, enabling other teams to reuse the benchmark.", "scoring": "Metrics include worse-cited wins, correctness-conflict accuracy, and Tie resolution across prompt strengths, with human validation separating Tie meanings.", "task": "Audit citation-like artifacts in RAG and agent-workflow evaluation through seven diagnostic conditions."}, "model": "deepseek-v4-pro", "reviewedAt": "2026-09-19T08:55:40.500135Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "A diagnostic benchmark for auditing citation-like artifacts in retrieval-augmented generation and agent-workflow evaluation, including content-equivalent pairs, citation ablations, correctness conflicts, quality gaps, prompt-strength ladders, decoupled judging, and workflow ranking.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence"], "snippet": "We introduce TraceJudgeBench, a diagnostic benchmark for auditing citation-like artifacts in RAG and agent-workflow evaluation, covering content-equivalent pairs, citation ablations, correctness conflicts, human-validated soft and moderate quality gaps, prompt-strength ladders, decoupled judging, and a controlled workflow-ranking probe."}, "familyId": "bmf_8668b2bc1d30", "firstSeenAt": "2026-09-14", "heat": null, "id": "bm_tracejudgebench_6d9b999a", "indexedAt": "2026-09-14T00:39:56.575498Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2609.12439", "pdf": "https://arxiv.org/pdf/2609.12439", "project": null, "report": "https://arxiv.org/abs/2609.12439"}, "metrics": [], "motivation": "LLM-as-a-judge protocols are commonly debiased by instructing judges to ignore presentation cues such as citation formatting, source labels, and evidence-display style.", "name": "TraceJudgeBench", "oneLine": "A diagnostic benchmark for auditing citation-like artifacts in retrieval-augmented generation and agent-workflow evaluation, including content-equivalent pairs, citation ablations, correctness conflicts, quality gaps, prompt-strength ladders, decoupled judging, and workflow ranking.", "paperTitle": "Debiasing as a Measurement Intervention: Calibrated Ties and Resolution Loss in LLM-as-a-Judge Evaluation", "primaryDomain": "General AI", "publication": {"evidence": "Accepted at Emnlp 2026", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2609.12439", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "Emnlp 2026", "verifiedAt": "2026-09-14T00:39:56.575498Z"}, "publications": [], "ranking": {"30d": {"confidence": "Low", "coverage": 0.0, "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": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}, "90d": {"confidence": "Low", "coverage": 0.0, "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": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}}, "readiness": "Paper only", "recognitionConfidence": 0.85, "relation": "introduces", "releasedAt": "2026-09-11", "source": {"authors": ["Liang Zhao and Yong Wang and Jiangzhe Chen"], "categories": ["cs.DL", "cs.AI"], "id": "2609.12439", "title": "Debiasing as a Measurement Intervention: Calibrated Ties and Resolution Loss in LLM-as-a-Judge Evaluation", "type": "arxiv", "url": "https://arxiv.org/abs/2609.12439"}, "sourceUpdatedAt": "2026-09-11T04:59:10Z", "topics": ["Agents"], "venueAttempts": [{"decisionRaw": "Accepted at Emnlp 2026", "evidence": [{"level": "author-claim", "observedAt": "2026-09-14T00:39:56.575498Z", "rawValue": "Accepted at Emnlp 2026", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2609.12439"}], "reviewStatus": "accepted", "venueName": "Emnlp 2026"}], "whyItMatters": "Addresses the lack of robust evaluation for LLM-as-a-judge artifacts, potentially improving the reliability of automated quality assessments."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
