Unverified50% confidenceFactExact time
Training-time alignment (such as RLHF and DPO), which is often contrasted with inference-time alignment, writes value preferences directly into the weights, offering greater stability but poor flexibility, and making it difficult to dynamically adjust for individual users.
1
Sources
50%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the Agent context, reinforcement learning (RL) is mainly used to optimize decision-making policies, with commonly used methods including RLHF and DPO; compared with supervised fine-tuning, it can explore better paths through trial and error, but it faces challenges such as difficulty in reward design, low sample efficiency, and unstable training.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74% similarUnverifiedCapability guardrails typically include value alignment during RLHF training, content filtering at the inference layer, and legal constraints from usage policies.74% similarUnverifiedThe sycophancy problem refers to models reinforcing accommodating behavior during RLHF training because responses that satisfy users receive higher human ratings, even at the cost of accuracy or safety.74% 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.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1038541API
curl https://kongchang.com/api/v1/knowledge/claims/1038541MCP
get_claim(id=1038541)