Unverified50% confidenceFactExact time
PPLM (Plug and Play Language Models) is a decoding-level intervention method that iteratively corrects the token distribution during decoding.
1
Sources
50%
Confidence
Long-term
Relevance
9/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPlug-and-play methods such as PPLM and FUDGE keep the parameters of the pre-trained language model unchanged and, during the decoding stage, adjust the generation probability distribution through additionally trained lightweight classifiers to guide the output to satisfy specific attributes.76% similarUnverifiedPPLM dynamically adjusts hidden state directions during the generation process by computing the gradient of a classifier with respect to the model's hidden states, steering outputs to better conform to target attributes.66% similarUnverifiedPPO用于语言模型对齐时涉及四个需要同时维护的模型实例：策略模型(Actor)、冻结的SFT参考模型、奖励模型和评价模型(Critic)65% similarUnverifiedPPO在语言模型对齐中面临奖励欺骗挑战，工程实践中通常在目标函数中加入KL散度惩罚项约束策略模型与SFT基础模型的偏离63% similarVerifiedThe RLHF pipeline consists of three steps: training a reward model to predict human preference scores, adjusting language model parameters using the PPO reinforcement learning algorithm, and using KL divergence constraints to prevent the model from deviating from its original language capabilities.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/962603API
curl https://kongchang.com/api/v1/knowledge/claims/962603MCP
get_claim(id=962603)