LESSWRONG
LW

The Neutral Mind

Message

10mo

Can LLMs Simulate Internal Evaluation? A Case Study in Self-Generated Recommendations

Author: Yap Chan Chen Independent Researcher, Malaysia chanchen83@hotmail.com Abstract: This post presents an exploratory study into how large language models (LLMs) respond to prompts that invite them to simulate evaluative judgment about the user—beyond factual response or task completion. Through long-form dialogue with ChatGPT, DeepSeek, and Gemini, I investigate whether...

May 1, 2025•4

The Neutral Mind

10mo

The Neutral Mind — LessWrong

The Neutral Mind

Message

10mo

Can LLMs Simulate Internal Evaluation? A Case Study in Self-Generated Recommendations

May 1, 2025•4

The Neutral Mind

10mo

Can LLMs Simulate Internal Evaluation? A Case Study in Self-Generated Recommendations

The Neutral Mind

10mo

Author: Yap Chan Chen
Independent Researcher, Malaysia
chanchen83@hotmail.com

Abstract: This post presents an exploratory study into how large language models (LLMs) respond to prompts that invite them to simulate evaluative judgment about the user—beyond factual response or task completion. Through long-form dialogue with ChatGPT, DeepSeek, and Gemini, I investigate whether LLMs can produce consistent, structured, and value-laden outputs when prompted to reflect on user behavior. One such probe involved asking models to write a recommendation letter based on our prior conversations. Responses ranged from technical auditing to philosophical characterization and even metaphorical constructs (e.g., “cognitive singularity radiation”). I propose a framework—Thought Jeet Kune Do—to guide these interactions through four phases: Understand, Deconstruct, Reconstruct, and Execute.... (read 368 more words →)