HF RL Explorer

Finding a target probability distribution for LLM confidence metrics:

Finding a target probability distribution for LLM confidence metrics:: a task in terminal-bench-2.0 (Harbor dataset). The confidence of a token prediction in an LLM can be quantified using different metrics. This implementation focuses on two metrics based on KL divergence from the uniform…

The task

The confidence of a token prediction in an LLM can be quantified using different metrics. This implementation focuses on two metrics based on KL divergence from the uniform distribution:

Part of introvoyz041/terminal-bench-2.0.