Unverified50% confidenceFactExact time
DreamTrue's reward model training dataset covers three categories of problems—robot defects (such as joint anomalies, self-intersection), object defects (such as objects vanishing into thin air or deforming incorrectly) and interaction defects (such as misaligned hand-object contact points, penetration)—and has undergone manual annotation.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified过拟合是LoRA训练中最常见的失败模式，模型会死记硬背训练图像像素特征而非学习本质特征68% similarVerifiedReinforcement learning acquires locomotion skills through massive trial-and-error in simulation environments, with reward functions that include metrics such as energy efficiency, driving robot movements toward being effortless and fluid, thereby producing human-like characteristics.65% similarUnverifiedThe multi-stage training of the AI alignment problem can introduce risks at every stage, including pre-training data bias, subjective annotator bias, imperfect approximation of the reward model, and excessive PPO reward optimization.65% similarUnverified排查模型训练失败应按照数据管线、训练过程、评估指标三层顺序逐步定位问题64% similarUnverifiedCapability evaluation methods fall into three categories: Benchmarks, Scenario-based Evals, and Human-AI Uplift Evals.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1012821API
curl https://kongchang.com/api/v1/knowledge/claims/1012821MCP
get_claim(id=1012821)