Safety
Think-Probe-Respond: Improving Large Language Models as Judges of Research Idea Novelty
arXiv:2608.25660v1 Announce Type: new Abstract: Automated novelty judgment can accelerate scientific discovery by enabling efficient evaluation, refinement, and comparison of research ideas. While lar
arXiv:2608.25660v1 Announce Type: new Abstract: Automated novelty judgment can accelerate scientific discovery by enabling efficient evaluation, refinement, and comparison of research ideas. While large language models are increasingly adopted for this task, we investigate a previously overlooked limitation in their judgment capabilities: despite generating reasoning rationales that closely mirror those of human experts, their final novelty judgments often diverge substantially. We demonstrate that this miscalibration stems from a systematic bias towards judging ideas as "medium novel". To mitigate this, we propose Think-Probe-Respond (TPR), a lightweight approach that probes latent novelty judgments from hidden states during the reasoning phase and uses the probed judgments to condition the final response. Across strong baselines, TPR improves novelty judgment performance by 22.30% and successfully mitigates the prevalent "medium novelty" bias.
Related
- DR-Arena: an Automated Evaluation Framework for Deep Research Agents
- Multi-Task Reinforcement Learning for Enhanced Multimodal LLM-as-a-Judge
- Do Methods Support the Claims? Intra-Paper Verification for Peer Review
- Faithful or Fabricated? A Causal Framework for Rationalization Bias in LLM Judges
Source: arXiv cs.CL | 2026-08-27