{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.03071", "kind": "paper", "url": "https://arxiv.org/abs/2608.03071"}], "authors": ["Guoyao Yu", "Xiaoqing Sun", "Ziqi Huang", "Shaojing Fan", "Zhongyi Zhang", "Xiaomeng Hu", "Xiaobo Xue", "Yangyang Shi", "Xiong Xiao", "Yang Song", "Biao Lyu", "Rong Wen", "Xing Li", "Qinming He", "Shunming Zhu", "Zhenguang Liu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.03071", "source_url": "https://arxiv.org/abs/2608.03071"}, "categories": ["General AI", "Language & Knowledge", "Reasoning"], "description": {"en": "ParamBench is a benchmark for evaluating LLM tool call parameter generation, built from real cloud-network APIs with difficulty tiers and exact match metrics. It is used to evaluate the proposed probe-guided training framework."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.03071", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.03071", "title": "ParamBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.03071", "modality": null, "name": "ParamBench", "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_parambench_25e30479", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "abd845006fb4521f76ab2e1b005b58084c3050199d42ebddff00e1ff9992fdec", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2608.03071"}, "publisher": null, "released": "2026-08-04", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.03071", "source_url": "https://arxiv.org/abs/2608.03071"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-03071", "source": "claire_radar", "source_benchmark_id": "2608.03071", "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/2608.03071", "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"}, "capabilities": ["Reasoning"], "confidence": "High", "construction": "Transform Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "86610886987ac92dc1bf815d5db268a7fa427bc436c85d171ce3f15d019d96de", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "ParamBench is a benchmark for evaluating LLM tool call parameter generation, built from real cloud-network APIs with difficulty tiers and exact match metrics. It is used to evaluate the proposed probe-guided training framework.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact named benchmark artifact released in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "To support systematic evaluation, we release ParamBench, a benchmark built from real cloud-network APIs that categorizes every instance into five difficulty levels according to parameter nesting depth, cross-parameter dependencies, and the reasoning required to derive values from earlier calls."}, "familyId": "bmf_00f78109a13a", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_parambench_25e30479", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2608.03071", "paper": "https://arxiv.org/abs/2608.03071", "pdf": "https://arxiv.org/pdf/2608.03071", "project": null, "report": "https://arxiv.org/abs/2608.03071"}, "metrics": [], "motivation": "Large language model agents derive much of their capability from tool use.", "name": "ParamBench", "oneLine": "ParamBench is a benchmark for evaluating LLM tool call parameter generation, built from real cloud-network APIs with difficulty tiers and exact match metrics. It is used to evaluate the proposed probe-guided training framework.", "paperTitle": "Getting the Parameters Right: A Difficulty-Graded Benchmark and Probe-Guided Training for LLM Tool Calls", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.03071", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {"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": 1.0, "relation": "introduces", "releasedAt": "2026-08-04", "source": {"authors": ["Guoyao Yu", "Xiaoqing Sun", "Ziqi Huang", "Shaojing Fan", "Zhongyi Zhang", "Xiaomeng Hu", "Xiaobo Xue", "Yangyang Shi", "Xiong Xiao", "Yang Song", "Biao Lyu", "Rong Wen", "Xing Li", "Qinming He", "Shunming Zhu", "Zhenguang Liu"], "categories": ["cs.AI"], "id": "2608.03071", "title": "Getting the Parameters Right: A Difficulty-Graded Benchmark and Probe-Guided Training for LLM Tool Calls", "type": "arxiv", "url": "https://arxiv.org/abs/2608.03071"}, "sourceUpdatedAt": "2026-08-04T03:36:41Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "Tool call parameter correctness is critical for execution yet understudied. ParamBench provides a systematic evaluation of parameter generation across difficulty levels, showing large improvements from probe-guided methods."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
