按年被提到
20242025 · 1 次2026
大模型与语言模型›提示工程与输出稳定性·arXiv:2408.04667 · cs.CL

Non-Determinism of "Deterministic" LLM Settings

Berk Atil, Sarp Aykent, Alexa Chittams, Lisheng Fu, Rebecca J. Passonneau, Evan Radcliffe, Guru Rajan Rajagopal, Adam Sloan, Tomasz Tudrej, Ferhan Ture, Zhe Wu, Lixinyu Xu 等 13 人 · 2024
开始阅读 ↗PDFalphaXiv 讨论
被提及 · 2 人首次 2025-11-04 · 最近 2026-04-09
2不同的人
2活跃月份
2讨论串
114被引(OpenAlex)
arXiv 摘要(候选层不写导读)

LLM (large language model) practitioners commonly notice that outputs can vary for the same inputs under settings expected to be deterministic. Yet the questions of how pervasive this is, and with what impact on results, have not to our knowledge been systematically investigated. We investigate non-determinism in five LLMs configured to be deterministic when applied to eight common tasks in across 10 runs, in both zero-shot and few-shot settings. We see accuracy variations up to 15% across naturally occurring runs with a gap of best possible performance to worst possible performance up to 70%. In fact, none of the LLMs consistently delivers repeatable accuracy across all tasks, much less identical output strings. Sharing preliminary results with insiders has revealed that non-determinism perhaps essential to the efficient use of compute resources via co-mingled data in input buffers so this issue is not going away anytime soon. To better quantify our observations, we introduce metrics focused on quantifying determinism, TARr@N for the total agreement rate at N runs over raw output, and TARa@N for total agreement rate of parsed-out answers. Our code and data are publicly available at https://github.com/breckbaldwin/llm-stability.

候选层:至少 2 个不同的人、在 2 个不同月份提到过,还没到主书架的 3 人 3 个月。没有做立场标注,所以只写“被提及”。

HN 书评

推荐在前,同类按时间 · 每篇最多保留 20 条,统计用全部提及
评论District5524 · 2025-11

“Not that it's incorrect but there is some data showing variability even with the very same input and all parameters. Especially if we have no control over the model behind the API with engineering optimizations etc. See Berk Atil et al.: Non-Determinism of "Deterministic" LLM Settings, https://arxiv.org/abs/2408.04667v5”

Lessons from interviews on deploying AI Agents in production原评论 ↗
评论yunohn · 2026-04

“I initially thought the same, but apparently with the inaccuracies inherent to floating-point arithmetic and various other such accuracy leakage, it’s not true! https://arxiv.org/html/2408.04667v5”

Claude mixes up who said what and that's not OK原评论 ↗