cs.AINov 3, 2026

Feeling of Knowing in Large Language Models

Authors: Zichuan Fu, Xian Wu, Jingtong Gao, Wenlin Zhang, Jiaxuan Li, Binhao Wang, Yimin Deng, Guojing Li, 

Organizations: City University of Hong Kong · Tencent Jarvis Lab · Xi’an Jiaotong University · Renmin University of China · 

Abstract

Post-training updates change what large language models (LLMs) can answer, but may leave their answerability judgments misaligned. We study this knowledge-judgment desynchronization through functional feeling of knowing (FoK), which supports selective generation, retrieval triggering, and model routing without implying subjective experience. We propose two trainable channels: a knowledge channel that changes answerability and a FoK channel that judges the current knowledge state. Three stages train them: on-policy supervised fine-tuning (OP-SFT) updates knowledge from corrected on-policy answers; knowledge-delta FoK training (KΔΔ-FT) adapts judgments across intermediate knowledge states; and FoK-guided policy optimization (FGPO) trains the subsequent answer or explanation. Factual-QA experiments, cross-dataset evaluation, and three sequential updates show that explicit FoK alignment improves judgment quality on updated knowledge. MMLU and ARC-Easy separately assess capability preservation.

Figures & tables

Appendix figures & tables27 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. What Models Know, How Well They Know It: Knowledge-Weighted Fine-Tuning for Learning When to Say "I Don't Know"

    Apr 7, 2026Joosung Lee, Hwiyeol Jo, Donghyeon Ko +3LLM AbstentionFine-Tuning

  2. GRADE: Probing Knowledge Gaps in LLMs through Gradient Subspace Dynamics

    Apr 3, 2026Yujing Wang, Yuanbang Liang, Yukun Lai +2LLM InterpretabilityLLM Uncertainty Estimation

  3. Beyond Questions: Evaluating the Stated Knowledge of LLMs

    May 26, 2026Luca Giordano, Simon RazniewskiLLM EvaluationFactual Knowledge in Language Models