强制JSON格式让44个模型答案趋同，削弱多样性。
Structured Output Collapses Answer Diversity Across 44 Language Models
- 仅要求返回JSON格式，不加约束，就导致答案趋同。
- 平均答案困惑度下降0.22比特，独特答案从52个减至36个。
- 适合关注模型输出一致性的研究者和部署工程师。
当语言模型需在大量合理选项中选择单一答案时，仅要求以JSON格式回复（无模式校验、无解码约束），就会显著改变其选择结果。我们重运行了One-Word Census实验：31个大答案空间类别提示，覆盖44个模型，新增仅要求返回JSON的条件。结果显示：在“选一个词”任务中，最常见答案占比从41%升至64%，独特答案数由52降至36，平均答案困惑度从1.80比特降至1.58比特。六种模型个体发生显著趋同（BH-FDR q=0.10），尤其原本最具个性的模型更明显，而保守模型几乎不变。重新采样显示，53%的模型稳定聊天默认回答在JSON格式下被替换为群体主流答案；部分模型如Claude Fable 5在聊天中从不答“cerulean”，但在JSON中达100%。全量控制实验揭示：压缩效应仅存在于模型训练时习惯的格式（JSON -0.22比特，p=0.0002；XML -0.19，p=0.002），YAML与CSV无影响，任意括号包裹甚至反向增加（+0.13，p=0.009），表明机制指向工具使用后训练的响应偏好。强制解码层校验（response_format）未带来额外压缩（-0.03比特），说明收敛发生在模型对格式的响应层面，而非解码过程。结构化输出是软件调用模型的接口，但该接口暴露的模型行为比聊天界面更均质，影响模型评估与选型。
原文摘要 · Abstract (English)
When a language model must choose one answer from a large space of equally valid options, a format clause -- "Reply with JSON only" -- changes which answer it chooses. We re-run the One-Word Census (arXiv:2607.12796): 31 wide-answer-space category prompts asked of 44 models, now with the reply requested in JSON -- no schema enforcement, no constrained decoding, only the request. Convergence deepens sharply: on the unconstrained "Pick a word" prompt the modal answer rises from 41% to 64% of the pool and distinct answers fall from 52 to 36; mean answer-choice surprisal drops from 1.80 to 1.58 bits. The tax is progressive: six of 44 models move individually (BH-FDR q=.10), all toward the mode, led by the most distinctive models, while the conformist floor is immobile. It is a sharpener, not a re-indexer -- the plain-chat modal answer survives in 28 of 31 categories. Defaults are register-indexed: a within-run re-sample (n=20) finds JSON shifts 53% of a model's stable chat defaults, mostly back to the crowd, and installs defaults absent from chat (Claude Fable 5 answers "cerulean" for colour 0% of the time in chat, 100% in JSON). Full-battery controls reveal a register gradient: compression is significant and specific to the answer-delivery formats models are trained to speak (JSON -0.22 bits, p=.0002; XML -0.19, p=.002), absent for YAML and CSV, and reversed for an arbitrary bracket wrapper (+0.13, p=.009) -- weighing the mechanism toward tool-use post-training. Enforcing the schema at the decoder (response_format) compresses no further than the request (-0.03 bits): the collapse lives in the model's response to the register, not the decoder. Structured output is how software consumes language models, and that surface is served by a measurably more homogeneous model than the chat surface on which models are evaluated, compared, and chosen.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。