Can I read Lexical Ambiguity in LLMs Study on EtoBox?
Lexical Ambiguity in LLMs Study by Myra Arora is a document available to read on EtoBox.
What is Lexical Ambiguity in LLMs Study about?
The document explores the impact of lexical ambiguity on the performance of large language models (LLMs) like GPT-4, Claude, and Gemini, highlighting how ambiguous prompts can lead to decreased accuracy and increased hallucinations. A study with 40 prompt pairs revealed that all models struggled with ambiguity, with Gemini performing the best overall. The findings suggest the need for improved ambiguity detection and confidence calibration in LLMs to enhance their reliability.
- Author
- Myra Arora
- Language
- EN