Improving In-Context Learning with Prediction Feedback for Sentiment Analysis
This work addresses sentiment analysis for users of LLMs, offering an incremental improvement over conventional methods.
The paper tackled the challenge of large language models distinguishing subtle sentiments in sentiment analysis by enhancing in-context learning with prediction feedback, resulting in an average F1 improvement of 5.95% across nine datasets.
Large language models (LLMs) have achieved promising results in sentiment analysis through the in-context learning (ICL) paradigm. However, their ability to distinguish subtle sentiments still remains a challenge. Inspired by the human ability to adjust understanding via feedback, this paper enhances ICL by incorporating prior predictions and feedback, aiming to rectify sentiment misinterpretation of LLMs. Specifically, the proposed framework consists of three steps: (1) acquiring prior predictions of LLMs, (2) devising predictive feedback based on correctness, and (3) leveraging a feedback-driven prompt to refine sentiment understanding. Experimental results across nine sentiment analysis datasets demonstrate the superiority of our framework over conventional ICL methods, with an average F1 improvement of 5.95%.