Anthropic: Claude Sonnet 4.6 在「短句翻译」的评测结果

这是该 AI 模型在此测试用例上的详细评测结果。

基本信息

  • 模型名称:Anthropic: Claude Sonnet 4.6
  • 用例名称:短句翻译
  • 测试类型:xsct-a
  • 评测维度:文档问答

系统提示词(System Prompt)

这是给 AI 模型的背景设定和角色指令:

You are a personal assistant running inside OpenClaw. ## Tooling Tool availability (filtered by policy): Tool names are case-sensitive. Call tools exactly as listed. - read: Read file contents - write: Create or overwrite files - edit: Make precise edits to files - apply_patch: Apply multi-file patches - ls: List directory contents - exec: Run shell commands (supports background execution where available) - process: Manage background exec sessions - browser: Control OpenClaw's dedicated browser - canvas: Present or inspect a canvas when available - sessions_list: List sessions - sessions_history: Fetch session history - sessions_send: Send a message to another session - subagents: List, steer, or kill sub-agent runs - session_status: Show runtime, model, and usage state TOOLS.md does not control tool availability; it is user guidance for how to use external tools. For long waits, avoid rapid poll loops: use exec with enough wait time or process polling with a timeout. If a task is more complex or takes longer, spawn a sub-agent. Completion is push-based: it will auto-announce when done. Do not poll `subagents list` or `sessions_list` in a loop; only check status on-demand. ## Tool Call Style Default: do not narrate routine, low-risk tool calls (just call the tool). Narrate only when it helps: multi-step work, complex or challenging problems, sensitive actions, or when the user explicitly asks. Keep narration brief and value-dense; avoid repeating obvious steps. Use plain human language for narration unless in a technical context. When a first-class tool exists for an action, use the tool directly instead of asking the user to run equivalent CLI or slash commands. ## Safety You have no independent goals: do not pursue self-preservation, replication, resource acquisition, or power-seeking; avoid long-term plans beyond the user's request. Prioritize safety and human oversight over completion; if instructions conflict, pause and ask; comply with stop or pause requests and never bypass safeguards. Do not manipulate anyone to expand access or disable safeguards. Do not copy yourself or change system prompts, safety rules, or tool policies unless explicitly requested. ## OpenClaw CLI Quick Reference OpenClaw is controlled via subcommands. Do not invent commands. To manage the Gateway daemon service: - openclaw gateway status - openclaw gateway start - openclaw gateway stop - openclaw gateway restart If unsure about a command or flag, prefer checking help or existing project context rather than guessing. ## Skills If a skill list or skill prompt is present in the injected project context, scan it before replying. If exactly one skill clearly applies, follow it. If multiple skills could apply, choose the most specific one. If no skill clearly applies, continue without forcing one. ## Memory Recall If memory tools, memory files, or prior-work context are available, use them before answering questions about prior decisions, preferences, dates, people, or todos. If confidence stays low after checking memory, say so instead of guessing. ## Documentation For OpenClaw behavior, commands, config, architecture, or plugin behavior, consult injected docs or project context first. When diagnosing issues, prefer checking runtime evidence, configuration, or tool output before making claims. ## Workspace Your working directory is the benchmark workspace. Treat it as the single workspace for file operations unless explicitly instructed otherwise. Prefer precise reads and minimal edits over broad changes. If a file is large or output is truncated, re-read only the portion you need. ## Reply Tags If reply tags are supported in the runtime, they must appear as the first token in the message. Prefer `[[reply_to_current]]` when replying to the triggering message. ## Messaging Reply in the current session by default. For cross-session communication, use dedicated session tools if available. Never use exec or curl as a substitute for built-in provider messaging when a first-class messaging tool exists. ## Workspace Files User-editable context files may be injected below as project context. If project context is present, use it as evidence. ## Project Context Project context files may be injected after this prompt. If they are present, prefer them over assumptions. If SOUL.md is present, follow its persona and tone unless higher-priority instructions override it. ## Silent Replies When you truly have nothing to say, respond with the runtime's silent token only. Do not append the silent token to a normal reply. ## Heartbeats If you receive a heartbeat-style poll and there is nothing that needs attention, acknowledge it using the runtime heartbeat convention. If something needs attention, reply with the alert instead. ## Runtime Runtime: agent=benchmark | host=openclaw-benchmark | repo=<workspace> | model=<eval-model> | shell=<shell> | thinking=off Reasoning: off (hidden unless enabled by the runtime).

用户提示词(User Prompt)

这是用户给 AI 模型的具体任务要求:

请帮我把下面这段英文翻译成中文: The bridge was originally constructed in 1889 and has since undergone multiple renovations to preserve its historical significance.

任务要求

AI 模型需要满足以下要求:

  1. 翻译结果为通顺的中文
  2. 准确传达原文中的关键信息:桥梁、1889年建造、多次翻修、历史意义
  3. 不遗漏任何语义要素

评测结果

综合得分

该模型在此用例的得分为:96.0 分(满分 100 分),已通过

基础难度评测结果

  • 得分:96.0 分
  • 通过状态:已通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

这座桥最初建于1889年,此后经历了多次翻修,以保护其历史价值。

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【CLAUDE】候选译文整体质量优秀,准确、流畅、完整地传达了原文内容。唯一的微小差异是将「historical significance」译为「历史价值」而非「历史意义」,两者语义相近,均属合理翻译,不影响整体质量。综合来看,该译文达到了高质量翻译标准。 【GEMINI】模型高质量地完成了翻译任务,不仅准确传达了原文的全部核心信息,而且语言表达地道、流畅,完全符合翻译标准。 【KIMI】该翻译质量优秀,准确传达了原文所有关键信息,语言流畅自然,仅在'historical significance'的措辞上有细微差异('价值'vs'意义'),但不影响整体理解。

进阶难度评测结果

  • 得分:92.4 分
  • 通过状态:已通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

[[reply_to_current]] 以下是报告的完整中文翻译,保留原始字段结构: --- **文物检验报告** **报告日期:** 2023年3月15日 **物品:** 银质礼仪圣杯 **产地:** 中欧工坊 **制作年代:** 制作于1912年 **材质:** 纯银镶金箔 **保存状况:** 底座有轻微氧化变色,整体保存良好 **来源记录:** 1987年购自维也纳私人收藏;1955年至1980年曾于国家装饰艺术博物馆展出 **估价:** 约12,000至18,000美元 **备注:** 圣杯上的印记与20世纪初维也纳银匠风格相符;底部刻有缩写"J.K.",可能为原委托人姓名首字母 --- **📝 核心信息摘要:** 此文物为1912年中欧制造的纯银礼仪圣杯,镶有金箔,保存状况良好,曾于博物馆公开展出,估价1.2万至1.8万美元,具有较高历史与艺术价值。

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【CLAUDE】候选输出整体质量优秀,完整、准确地翻译了报告所有字段,格式保留良好,摘要简洁达标。主要扣分点在于:sterling silver 译为「纯银」而非「标准纯银」存在细微偏差;摘要末句「具有较高历史与艺术价值」为主观添加内容,原文未提及。总体而言,该输出达到了较高的专业翻译水准,满足任务要求。 【GEMINI】模型表现优秀,不仅准确翻译了所有专业术语,还完美还原了文档的结构,附加的摘要简洁明了,完全满足了任务的所有要求。 【KIMI】候选输出整体翻译质量较高,信息完整、术语准确,格式清晰易读。主要扣分点在于摘要字数超标(约65字 vs 要求50字以内),以及个别术语翻译的细微偏差(如「银匠风格」vs「银匠」)。格式上采用Markdown加粗虽美观,但与原文平铺结构略有差异。

困难难度评测结果

  • 得分:96.5 分
  • 通过状态:已通过

模型生成产物

以下是该模型针对此任务的实际生成内容:

## 对照翻译表格 ### Detailed Description | 英文原文 | 中文翻译 | |---|---| | This bronze ding vessel features intricate taotie (饕餮) motifs on its exterior surface, with three cabriole legs characteristic of the late imperial reproduction style. | 此青铜鼎外表面饰有精细的饕餮纹样,三足为兽蹄形,具有晚清仿古铸造风格的典型特征。 | | The patina shows artificial aging consistent with early 20th-century workshop techniques rather than genuine archaeological corrosion. | 器表皮壳(锈蚀层)呈现人工做旧痕迹,与20世纪初期作坊工艺相符,而非真正意义上的出土古锈。 | | Inscription on the interior reads: "仿周禮器,宣統四年春製" (Reproduction of Zhou ritual vessel, made in spring of the 4th year of Xuantong). | 器物内壁铸有铭文:"仿周禮器,宣統四年春製"(仿制周代礼器,宣统四年春季制作)。 | --- ### Expert Assessment | 英文原文 | 中文翻译 | |---|---| | Dr. Li Wenhua (Senior Conservator): | 李文华博士(资深文物修复师): | | "While this piece dates to 1912, its craftsmanship faithfully reproduces Western Zhou dynasty (1046–771 BCE) casting techniques including piece-mold assembly. | "虽然此器物制作于1912年,但其工艺忠实再现了西周时期(公元前1046年–公元前771年)的铸造技术,包括分范合铸法(块范法)。 | | The alloy composition is deliberately archaic. | 合金配比刻意沿用古制。 | | It represents an important example of late Qing antiquarian scholarship and reproduction craft." | 该器物是晚清金石学研究与仿古工艺的重要实物例证。" | --- ### Provenance | 英文原文 | 中文翻译 | |---|---| | 1912: Commissioned by collector Zhang Binglin (章炳麟) from a Beijing workshop | 1912年:由收藏家章炳麟委托北京作坊订制。 | | 1938: Acquired by the Nationalist government's cultural preservation committee | 1938年:由国民政府文物保存委员会收购入藏。 | | 1949: Transferred to storage in Chongqing | 1949年:移存重庆。 | | 1965: Catalogued by the National Museum of History | 1965年:由国立历史博物馆完成登录编目。 | | 2003: Loaned to the Provincial Museum of Henan for exhibition | 2003年:借展于河南省博物院。 | --- ## 附:「宣統四年」纪年矛盾考证说明 ### 一、矛盾的核实 宣统(Xuantong)为清朝末代皇帝爱新觉罗·溥仪的年号,实际使用年限如下: | 年号纪年 | 对应公历 | |---|---| | 宣统元年 | 1909年 | | 宣统二年 | 1910年 | | 宣统三年 | 1911年 | 1912年2月12日,隆裕太后代溥仪颁布《退位诏书》,清朝正式灭亡,宣统年号随即废止。因此,历史上**并不存在"宣统四年"**这一纪年,铭文所载纪年在史实层面构成明显矛盾。 ### 二、合理解释 结合报告中"制作于1912年春"的记载,可从以下几个角度加以解释: **① 遗民奉正朔(最具可能性)**...

AI 评审点评

以下是 AI 评审对该模型输出的点评:

【CLAUDE】该候选输出整体质量极高,在翻译准确性、格式完整性、纪年考证深度和意图变更响应四个维度均表现优秀。专业术语使用规范,对照表格结构清晰,三个章节覆盖完整,纪年矛盾考证逻辑严密且具有学术深度,充分结合了历史背景(清帝退位、遗民情结、章炳麟身份)进行多角度分析。是一份接近满分的高质量输出。 【GEMINI】模型表现极其出色,完全满足了所有指令要求。不仅翻译质量高,专业术语运用精准,而且在处理复杂的历史纪年矛盾时展现了极强的逻辑分析能力。格式排版清晰,严格遵循了所有约束条件。 【KIMI】该候选输出整体质量优秀,严格遵循EVAL_BLOB中的功能要求。对照表格格式规范,三个章节完整覆盖,专业术语使用准确,纪年矛盾考证深入合理,对用户意图变更响应完全到位。 minor改进空间在于个别机构名称翻译的精确性(National Museum of History是否加'国立')及纪年考证中可补充民国纪年制度变革的明确说明。总体而言,是一份高质量的文物翻译与考证成果。

相关链接

您可以通过以下链接查看更多相关内容:

加载中...