metadata
license: mit
base_model:
- Klingspor/lta_singleturn_hard_sft_qwen3-4b
∆Belief-RL Post-trained Model
This model is a reinforcement learning (RL) post-trained language model optimized with the ∆Belief reward signal introduced in ∆Belief-RL.
Instead of relying solely on sparse success signals, the model is rewarded for reducing its own belief uncertainty over time. This enables dense feedback and turn-level credit assignment in long-horizon, open-ended information-seeking tasks.
The model was trained in the Twenty Questions environment and learns general information-seeking strategies that generalize beyond the training horizon.
Resources
- 🌐 Project page: https://bethgelab.github.io/delta-belief-rl/
- 💻 Code repository: https://github.com/bethgelab/delta-belief-rl