LLMs as Method Actors: A Model for Prompt Engineering and Architecture

We introduce “Method Actors” as a mental model for guiding LLM prompt engineering and prompt architecture. Under this mental model, LLMs should be thought of as actors; prompts as scripts and cues; and LLM responses as performances. We apply this mental model to the task of improving LLM performance at playing Connections, a New York Times word puzzle game that prior research identified as a challenging benchmark for evaluating LLM reasoning. Our experiments with GPT-4o show that a “Method Actors” approach can significantly improve LLM performance over both a vanilla and “Chain of Thoughts” approach. A vanilla approach solves 27% of Connections puzzles in our dataset and a “Chain of Thoughts” approach solves 41% of puzzles, whereas our strongest “Method Actor” approach solves 86% of puzzles. We also test OpenAI’s newest model designed specifically for complex reasoning tasks, o1-preview. When asked to solve a puzzle all at once, o1-preview solves 79% of Connections puzzles in our dataset, and when allowed to build puzzle solutions one guess at a time over multiple API calls, o1-preview solves 100% of the puzzles. Incorporating a “Method Actor” prompt architecture increases the percentage of puzzles that o1-preview solves perfectly from 76% to 87%.

Introduction. We introduce “Method Actors” as a mental model for guiding LLM prompt engineering and prompt architecture. Under this mental model, LLMs should be thought of as actors; prompts as scripts and cues; and LLM responses as performances. Four principles for prompt writing and task decomposition follow from this mental model: 1) Prompt engineering is playwriting and directing. 2) Performance re- We apply this mental model to the task of improving LLM performance at playing Connections, a New York Times word puzzle game that prior research has identified as a useful benchmark for testing LLM complex reasoning performance. (Samadarshi et al., 2024; Todd et al., 2024). With these puzzles, a player is shown a four-by-four grid of 16 words and must identify four groups of four words that have a unique connection to one another. Each word appears in exactly one of the four groups, and each of the four groups is unique. For each game, the player is allowed to make up to three incorrect guesses and still solve the puzzle.

Discussion / Conclusion. This paper demonstrates how using “Method Actors” as a mental model for LLMs can improve LLM performance with one particular complex reasoning task, but leaves open the opportunity to examine how this mental model affects LLM performance with different reasoning tasks or tasks distinct from reasoning, such as creative writing. There’s an opportunity to draw upon the acting literature to test whether methods for improving actors’ performances might also improve LLM performances. Just as method acting principles revolutionized acting in the mid-20th century to produce a new form of authentic onstage and onscreen performances, Although designing prompt architecture to mimic human structures of cognition may expand the reasoning abilities of LLMs, the field should not be confined to this approach. At times, it can be a useful analogy to think of LLM responses as thoughts.