Unverified50% confidenceFactExact time
In a 2025 analysis, Thinking Machines Lab named the vLLM temperature=0 nondeterminism problem "lack of batch invariance."
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
UnverifiedTemperature=0（贪心采样）也不能保证模型输出的确定性，至少有三个原因：浮点运算非结合性、MoE路由不确定性、服务端配置漂移68% similarUnverified即便温度参数设为0，不同的系统提示词、对话上下文和模型版本更新都会导致LLM输出差异，使品牌可见度测量面临可复现性挑战67% similarUnverifiedWhen an LLM scores as a judge, even with temperature set to 0, the same conversation trajectory and judge model run five times still produces a different verdict once, and this non-determinism is unacceptable for a test suite that wants to be integrated into CI.66% similarUnverifiedSetting the sampling temperature to 0 causes reasoning models to fall into a thinking loop and become stuck during the thinking phase.66% similarVerifiedThe non-determinism of large language models stems from their autoregressive sampling process; even with temperature set to 0, differences in inference frameworks, batch sizes, and hardware floating-point operations can all cause output drift.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1060908API
curl https://kongchang.com/api/v1/knowledge/claims/1060908MCP
get_claim(id=1060908)