{
  "abstract": "Large Language Models (LLMs) are increasingly deployed as scientific AI as- sistants, and a growing body of benchmarks evaluates their capabilities across knowledge retrieval, reasoning, code generation, and tool use. These evaluations, however, typically assume the scientific problem is already well-posed, whereas practical scientific assistance often begins with an ill-posed user request that must be refined through dialogue before any computation, analysis, or experiment can be carried out reliably. We introduce SCICONVBENCH, a benchmark for multi- turn clarification in scientific task formulation across four computational science problem domains: fluid mechanics, solid mechanics, materials science, and par- tial differential equations (PDEs). SCICONVBENCH targets two complementary capabilities: eliciting missing information (disambiguation) and detecting and correcting erroneous requests containing internally contradictory information (in- consistency resolution). Our benchmark pairs a structured task ontology with a rubric-based evaluation framework, enabling systematic measurement of LLM per- formance across three dimensions: clarification behavior, conversational grounding, and final-specification fidelity. Current frontier models perform relatively well on inconsistency resolution, but even the best model resolves only 52.7% of the disambiguation cases in fluid mechanics. We further find that frontier LLMs fre- quently make silent assumptions and perform implicit specification repairs that are not grounded in the conversation with users. SCICONVBENCH establishes a foundation for evaluating the upstream conversational reasoning that a reliable computational science assistant requires. The code and data can be found at https://github.com/csml-rpi/SciConvBench.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2605.18630",
  "author_total": 8,
  "authors": [
    {
      "name": "Nithin Somasekharan"
    },
    {
      "name": "Youssef Hassan"
    },
    {
      "name": "Shiyao Lin"
    },
    {
      "name": "Gihan Panapitiya"
    },
    {
      "name": "Patrick Emami"
    },
    {
      "name": "Anurag Acharya"
    },
    {
      "name": "Sameera Horawalavithana"
    },
    {
      "name": "Shaowu Pan"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/0cqqjfgsd64ha.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2605.18630",
  "fulltext_url": null,
  "id": "0cqqjfgsd64ha",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2605.18630",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.AI"
    ],
    [
      "arxiv",
      "physics.comp-ph"
    ],
    [
      "openalex",
      "1702|Speech and dialogue systems"
    ]
  ],
  "openalex": "W7161698529",
  "pdf_url": "https://arxiv.org/pdf/2605.18630",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-17",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/0cqqjfgsd64ha.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:22:19Z",
      "native_id": "2605.18630",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2605.18630"
    }
  ],
  "subjects": [
    {
      "field": "physics",
      "origin": "arxiv:physics.comp-ph",
      "subfield": "computational"
    }
  ],
  "title": "SCICONVBENCH: Benchmarking LLMs on Multi-Turn Clarification for Task Formulation in Computational Science",
  "title_text": "SCICONVBENCH: Benchmarking LLMs on Multi-Turn Clarification for Task Formulation in Computational Science",
  "type": "preprint",
  "updated_at": "2026-10-30T12:22:19Z",
  "url": "https://research.draconitic.com/paper/0cqqjfgsd64ha",
  "venue": "arXiv",
  "versions": []
}