open-ended-reasoning

ThinkProbe introduces a non-generative framework that converts LLM reasoning traces into thought graphs to derive a five-dimensional cognitive profile, revealing stable model-level reasoning structures and enabling evaluation beyond accuracy for open-ended tasks.

OPERA proposes a reinforcement learning method for open-ended tasks using intrinsic rewards based on perplexity dynamics, replacing unreliable LLM-as-a-judge reward models. It achieves state-of-the-art results on Qwen3-8B, matching proprietary models in creative writing and other open-ended tasks.