Language models improve when to say i dont know on factual questions

Can Language Models Learn When to Abstain? Policy Learning and Calibrated Answer Selection

Computation and Language

Summary

Knowing when to answer or say "I don't know" is important for language models. The authors tested two approaches: training the model itself or adding a separate answer checker to decide when to abstain. They found that training a separate answer selector helped language models avoid mistakes while answering more questions correctly on one dataset. However, this method was less effective on another dataset and the tested reinforcement learning did not show consistent improvements. Their work suggests answer selection can improve performance, but reliable abstention across different types of questions remains challenging.

What this means in practice

  • For ai model developers: Improve language model answer quality by adding learned components that decide when to abstain from uncertain answers on factual datasets.
  • For chatbot engineers: Develop chatbots that reduce errors by incorporating answer selectors to better decide when to provide answers or decline to respond.

Authors

Juan Terven, Diana-Margarita Córdova-Esparza, Ruben Gomez, Carlos Miranda Rocha, Julio-Alejandro Romero-González, Edgar Chavez-Urbiola

Abstract

A useful language model should know when to answer and when to say "I don't know." Can we improve that decision by training the model itself or by training a separate component to check its answers? We study both approaches on Qwen3-14B for factual questions, using a score that rewards correct answers, penalizes mistakes, and allows abstention. Across three training seeds and two held-out confidence thresholds, a learned answer selector improves threshold-averaged utility over the original model's calibrated confidence filter. On TriviaQA, mean coverage increases from 20.2% to 30.9%, while the pooled error rate among scored answers decreases from 23.0% to 14.5% under exact-match grading. The tested reinforcement-learning changes do not establish consistent gains. On NQ-Open, the selector improves threshold-averaged utility over the filter but still performs worse than always abstaining. A partial human-reviewed audit preserves the TriviaQA finding. Our results support learning to select answers in this setting while showing that reliable abstention across datasets remains unresolved.