paper-with-me

홈 › Papers

Aligning Language Models Using Follow-up Likelihood as Reward Signal

2024-09-20 · Chen Zhang, Dading Chong, Feng Jiang, Chengguang Tang, Anningzhe Gao, Guohua Tang, Haizhou Li

In natural human-to-human conversations, participants often receive feedback signals from one another based on their follow-up reactions. These reactions can include verbal responses, facial expressions, changes in emotional state, and other non-verbal cues. Similarly, in human-machine interactions, the machine can leverage the user's follow-up utterances as feedback signals to assess whether it has appropriately addressed the user's request. Therefore, we propose using the likelihood of follow-up utterances as rewards to differentiate preferred responses from less favored ones, without relying on human or commercial LLM-based preference annotations. Our proposed reward mechanism, ``Follow-up Likelihood as Reward" (FLR), matches the performance of strong reward models trained on large-scale human or GPT-4 annotated data on 8 pairwise-preference and 4 rating-based benchmarks. Building upon the FLR mechanism, we propose to automatically mine preference data from the online generations of a base policy model. The preference data are subsequently used to boost the helpfulness of the base model through direct alignment from preference (DAP) methods, such as direct preference optimization (DPO). Lastly, we demonstrate that fine-tuning the language model that provides follow-up likelihood with natural language feedback significantly enhances FLR's performance on reward modeling benchmarks and effectiveness in aligning the base policy model's helpfulness.

📄 PDF Abstract BibTeX arXiv:2409.13948

Code (1)

e0397123/flr 공식 구현 pytorch

Tasks

Language Modelling

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Linear Layer A Linear Layer is a projection $\mathbf{XW + b}$.
Position-Wise Feed-Forward Layer 설명 없음
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…
Absolute Position Encodings Absolute Position Encodings are a type of position embeddings for [Transformer-based models] where positional encodings are…
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…

Similar Papers 제목 키워드 기반

Reinforcement Learning from User Feedback

2025-05-20 · Eric Han, Jun Chen, Karthik Abinav Sankararaman, Xiaoliang Peng 외

As large language models (LLMs) are increasingly deployed in diverse user facing applications, aligning them with real user preferences becomes essential. Existing methods like Reinforcement Learning from Human Feedback …

reinforcement-learningReinforcement Learning

Label-Free Reinforcement Learning via Cross-Model Entropy

2026-05-27 · Matt Gorbett, Hossein Shirazi arxiv

Post-training large language models with reinforcement learning is bottlenecked by the reward signal. Existing approaches require either ground-truth verifiable rewards, restricting training to domains with automatic cor…

Reinforcement LearningInstruction Following

Discriminative Finetuning of Generative Large Language Models without Reward Models and Human Preference Data

2025-02-25 · Siqi Guo, Ilgee Hong, Vicente Balmaseda, Changlong Yu 외

Supervised fine-tuning (SFT) has become a crucial step for aligning pretrained large language models (LLMs) using supervised datasets of input-output pairs. However, despite being supervised, SFT is inherently limited by…

Aligning Dialogue Agents with Global Feedback via Large Language Model Reward Decomposition

2025-05-21 · Dong Won Lee, Hae Won Park, Cynthia Breazeal, Louis-Philippe Morency

We propose a large language model based reward decomposition framework for aligning dialogue agents using only a single session-level feedback signal. We leverage the reasoning capabilities of a frozen, pretrained large …

Dialogue GenerationLanguage ModelingLanguage ModellingLarge Language Model

Process Supervision of Confidence Margin for Calibrated LLM Reasoning

2026-04-25 · Liaoyaqi Wang, Chunsheng Zuo, William Jurayj, Benjamin Van Durme 외 arxiv

Scaling test-time computation with reinforcement learning (RL) has emerged as a reliable path to improve large language models (LLM) reasoning ability. Yet, outcome-based reward often incentivizes models to be overconfid…

Reinforcement Learning