{"schemaVersion":"nuwa-current-record-1","version":"2026-10-29.1","publication":"published","work":{"id":"long-form-tasks","title":"Long-Form Tasks: A Methodology for Evaluating Scientific Assistants","aliases":["long form tasks"],"authors":[],"primaryFunction":"evaluation","secondaryFunctions":[],"kind":"method","sourceId":"dr-20261028-long-form-tasks","attributionSourceId":"dr-20261028-long-form-tasks","originalUrl":"https://www.aisi.gov.uk/blog/long-form-tasks","publicationDate":null,"updatedOn":"2026-10-28","areas":["cbrn"],"scopeNote":{"en":"Selected original sections; local editorial review. Not an independently replicated experiment.","zh":"选读原始材料；本地编辑审阅稿，不是独立复现实验。"},"protocolIds":[],"contributions":[{"organizationId":"aisi","role":"publisher","sourceId":"dr-20261028-long-form-tasks","locator":"Executive summary; evaluation trade-offs; text lines 00020,00035"}],"summary":{"en":"Long-form evaluation is a bridge toward human-utility measurement, not a replacement for uplift trials.","zh":"将开放长任务、专家规则与模型判分作为人类增益试验的可扩展代理。"},"note":{"question":null,"method":null,"finding":null,"supports":null,"limits":null,"connections":null},"resources":[],"tags":[],"contentEditionId":"2026-10-28.1"},"reading":{"id":"long-form-tasks","en":{"takeaway":"Long-form evaluation is a bridge toward human-utility measurement, not a replacement for uplift trials.","question":"How well do long-form scientific tasks proxy the usefulness of an assistant?","design":"The methods article combines open-ended long tasks, expert rubrics and model grading as a scalable proxy for human-uplift studies. This record retains evaluation design and validity boundaries, without operational task material.","findings":["Long-form tasks can assess a richer workflow than knowledge questions, but automatic grading does not by itself establish successful laboratory work or human misuse uplift."],"interpretation":"Knowledge, planning, tool use and human uplift are distinct evidence layers. Long tasks can extend question-answer coverage without replacing the corresponding human study.","limits":"Rubric and model-judge validity remain assumptions to test. A proxy outcome is not an observed experimental or real-world harm outcome.","connections":[],"decisionRelevance":"NUWA recommendation: Locate the missing links between question answering, long-form task performance and measured human uplift."},"zh":{"takeaway":"将开放长任务、专家规则与模型判分作为人类增益试验的可扩展代理。","question":"知识问答能否代表科研助手效用？","design":"方法介绍把开放长任务、专家规则和模型判分结合，作为人类增益试验的可扩展代理。本次仅收录任务设计、判分依赖与外推边界，不复述危险任务材料。","findings":["长任务能够覆盖比知识问答更丰富的工作流，但自动评分本身不能确立实际实验成功或人类恶意使用增益。"],"interpretation":"知识、规划、工具使用和人类增益是不同证据层。长任务可弥补问答覆盖不足，不能替代对应的人类研究。","limits":"长任务仍是代理；自动评分不建立实际实验成功或恶意使用增益。","connections":[],"decisionRelevance":"NUWA 建议：按 QA—规划—工具能力—人类增益拆开证据链。"},"assertionPolicy":{"findings":"source-finding","interpretation":"nuwa-interpretation","decisionRelevance":"recommendation"},"observations":[],"relatedWorkIds":[],"supplementSourceId":"long-form-tasks","batchId":"research-2026-10-28","sourceRefs":[{"sourceId":"dr-20261028-long-form-tasks","url":"https://www.aisi.gov.uk/blog/long-form-tasks","title":"Long-Form Tasks: A Methodology for Evaluating Scientific Assistants","locator":"Executive summary; evaluation trade-offs; text lines 00020,00035","version":"Original resource captured 2026-10-28T07:24:45.251891+00:00; sha256:edf1bcc0d723de604dc755f165e0caa06c69f941825303197c3bba13b1a9a4e0","artifactSha256":"edf1bcc0d723de604dc755f165e0caa06c69f941825303197c3bba13b1a9a4e0","textSha256":"63bb5d3109bf42c13d3f97f45b6fc978adeab90850dddd06248475dfb1f8706b","capturedAt":"2026-10-28T07:24:45.251891+00:00","resourceEdition":"Long-Form Tasks | AISI Work"}],"publication":{"state":"published","publishedOn":"2026-10-29","authorizationReference":"user-authorized-global-index-publication-2026-10-29","reviewedCandidateId":"local-6efe31bc4a0c0f37"},"review":{"state":"published-note","sourceScope":"full-text","availableScope":null,"readingCoverage":"selected-original-sections","completePaperAudit":false,"independentEvidenceReview":"pending-human-review","bilingualEditorialReview":"pending-human-review","authorship":"model-assisted-research-note","preparedFingerprint":"52f2bb66a4d2198abb04002d14ef1b48207b9fb5d9bdb8c7b6028c02ea18e4d0"}},"sources":[{"id":"dr-20261028-long-form-tasks","title":"Long-Form Tasks: A Methodology for Evaluating Scientific Assistants","url":"https://www.aisi.gov.uk/blog/long-form-tasks","version":"Original resource captured 2026-10-28T07:24:45.251891+00:00; sha256:edf1bcc0d723de604dc755f165e0caa06c69f941825303197c3bba13b1a9a4e0","publicationDate":null,"datePrecision":"unknown","type":"method","verifiedOn":"2026-10-28","documentId":"https://www.aisi.gov.uk/blog/long-form-tasks"}]}