Unverified50% confidenceFactExact time
The experimental scale was 5 reward configurations × 5 random seeds × 2,000 episodes per run.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/7/2026
First Seen
Valid until: 12/6/2026
Sources
Related Claims
Unverified通过记录随机种子可精确复现导致失败的特定顺序，再通过二分法逐步缩小污染源的范围69% similarUnverifiedRandom seeds in simulation control the initial state of the pseudo-random number generator; different seeds produce different sequences of environmental perturbations and sensor noise samples, which is essentially an application of the Monte Carlo method.67% similarUnverified即便固定所有随机种子，在不同硬件或并行度下，训练结果仍可能出现被大模型放大为显著能力差异的细微偏差66% similarUnverified深度学习中的随机种子控制权重初始化、数据加载批次打乱顺序、Dropout神经元随机失活以及优化器中的随机梯度估计等多个随机性来源63% similarUnverifiedThe effect size Cohen's d between task performance under the naive survival reward configuration and a random policy was -10.6.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/871629API
curl https://kongchang.com/api/v1/knowledge/claims/871629MCP
get_claim(id=871629)