Correct-Detect: Balancing Performance and Ambiguity Through the Lens of Coreference Resolution in LLMs

Large Language Models (LLMs) are intended to reflect human linguistic competencies. But humans have access to a broad and embodied context, which is key in detecting and resolving linguistic ambiguities, even in isolated text spans. A foundational case of semantic ambiguity is found in the task of coreference resolution: how is a pronoun related to an earlier person mention? This capability is implicit in nearly every downstream task, and the presence of ambiguity at this level can alter performance significantly. We show that LLMs can achieve good performance with minimal prompting in both coreference disambiguation and the detection of ambiguity in coreference, however, they cannot do both at the same time. We present the CORRECT-DETECT trade-off: though models have both capabilities and deploy them implicitly, successful performance balancing these two abilities remains elusive. Unambiguous Kimberly told the aunt that she contacted the granddaughter. Who contacted the granddaughter? Ambiguous Kimberly told the aunt that the granddaughter trusted her. Who did the granddaughter trust? Ambiguous Kimberly told the aunt that the granddaughter trusted her. Who did the granddaughter trust? Ambiguous Mostly ambiguous Overdetects ambiguity Detect Ambig