human preferences
In AI, particularly in reinforcement learning and preference learning, human preferences refer to the incorporation of human judgments or choices into the training of models to align their behavior with what humans value or prefer in certain tasks or domains.
- BLEUBERI: BLEU is a surprisingly effective reward for instruction following
- Direct Alignment with Heterogeneous Preferences
- Distributive Fairness in Large Language Models: Evaluating Alignment with Human Values
- EmergentTTS-Eval: Evaluating TTS Models on Complex Prosodic, Expressiveness, and Linguistic Challenges Using Model-as-a-Judge
- InfiFPO: Implicit Model Fusion via Preference Optimization in Large Language Models
- Learning Preferences without Interaction for Cooperative AI: A Hybrid Offline-Online Approach
- On Evaluating LLM Alignment by Evaluating LLMs as Judges
- Pairwise Calibrated Rewards for Pluralistic Alignment
- Reducing the Probability of Undesirable Outputs in Language Models Using Probabilistic Inference
- Reverse Engineering Human Preferences with Reinforcement Learning
- Reward-Instruct: A Reward-Centric Approach to Fast Photo-Realistic Image Generation
- STAIR: Addressing Stage Misalignment through Temporal-Aligned Preference Reinforcement Learning
- VTON-VLLM: Aligning Virtual Try-On Models with Human Preferences
- WorldModelBench: Judging Video Generation Models As World Models