🎸 Emotion Prompting

Abstract

This section covers "Emotion Prompting".

🦜 Video lecture for this chapter - Link

✒️ Overview

Emotion prompting is a technique that involves adding emotional cues to prompts given to large language models (LLMs). An example of emotion cue is “This is very important to my career”.

✒️ How it works

This prompting approach is proposed based on previous psychological studies that have shown the beneficial impact of emotional stimuli related to expectancy, confidence, and social influence on individuals. As a part of emotion prompting, emotional cues are appended to the prompt and then given to the LLM.

From the results presented in the above image, emotion prompting enhances the results by a good margin in case of all the LLMs.

✒️ Pros

• Easy of Use - All you need to do is to append an emotion cue to the original prompt.
• Better Performance - Emotion prompting enhances the LLM performance by a good margin.

✒️ Cons

• Manipulation and bias - Malicious actors could use emotional prompting to manipulate LLMs into generating harmful content.
• Misinterpretation of cues - LLMs might misinterpret emotional prompts, leading to inaccurate or inappropriate outputs.

To summarize, emotion prompting is a simple prompting technique to enhance the performance of LLMs just by appending emotion cues.

Note: The above image is from "Emotion Prompting" Paper.