Unverified50% confidenceOpinionExact time
The researchers believe the issue is most likely a systemic byproduct of current mainstream reinforcement learning training paradigms (such as RLHF and RLEF), rather than an engineering error on the part of any individual vendor.
1
Sources
50%
Confidence
Long-term
Relevance
9/29/2026
First Seen
Sources
Related Entities
Related Claims
Verified当代前沿模型的训练方法包括RLHF（基于人类反馈的强化学习）及其后继者RLAIF和过程奖励模型（PRM）80% similarUnverifiedThe jev architecture used for comparison employs parallel sampling, is trained via RLCD (Reinforcement Learning from Contrastive Distillation), and outputs confidence distributions and schema selection.79% similarVerifiedTraditional RLHF requires first training a reward model and then optimizing with reinforcement learning algorithms such as PPO, making the process complex and unstable.78% similarUnverified近期研究表明RL（尤其是RLVR）并没有给模型注入大量新知识，而更像是激发和对齐预训练阶段已具备的能力77% similarUnverifiedRLHF和DPO等对齐技术的效果高度依赖预训练基础质量，预训练阶段未习得的能力几乎无法通过后续对齐补救77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/958409API
curl https://kongchang.com/api/v1/knowledge/claims/958409MCP
get_claim(id=958409)