Self-Feedback Reinforcement Learning
A reinforcement learning method in which a model uses its own outputs or a critic model to score responses in place of human annotation. Through preference optimization or policy gradient, it iteratively improves policies for long-horizon tasks.
Timeline (last 90 days)
DeepSeek officially stated that V4.1-Flash uses Self-Rewarding RL to make tasks such as search and fetching into training objectives.
DeepSeek says V4.1-Flash-1017 uses reinforcement learning with self-feedback to drive end-to-end project work, including game creation, 3D modeling and long agent runs.
All Facts (2)
DeepSeek officially stated that V4.1-Flash uses Self-Rewarding RL to make tasks such as search and fetching into training objectives.
50%UnverifiedDeepSeek says V4.1-Flash-1017 uses reinforcement learning with self-feedback to drive end-to-end project work, including game creation, 3D modeling and long agent runs.
50%