How FaR Are Large Language Models From Agents with Theory-of-Mind? Article Swipe

MAKE IMPACT

Pei Zhou , Aman Madaan , Srividya Pranavi Potharaju , Aditya Gupta , Kevin R. McKee , Ari Holtzman , Jay Pujara , Xiang Ren , Swaroop Mishra , Aida Nematzadeh , Shyam Upadhyay , Manaal Faruqui · YOU? · · 2023 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2310.03051

"Thinking is for Doing." Humans can infer other people's mental states from observations--an ability called Theory-of-Mind (ToM)--and subsequently act pragmatically on those inferences. Existing question answering benchmarks such as ToMi ask models questions to make inferences about beliefs of characters in a story, but do not test whether models can then use these inferences to guide their actions. We propose a new evaluation paradigm for large language models (LLMs): Thinking for Doing (T4D), which requires models to connect inferences about others' mental states to actions in social scenarios. Experiments on T4D demonstrate that LLMs such as GPT-4 and PaLM 2 seemingly excel at tracking characters' beliefs in stories, but they struggle to translate this capability into strategic action. Our analysis reveals the core challenge for LLMs lies in identifying the implicit inferences about mental states without being explicitly asked about as in ToMi, that lead to choosing the correct action in T4D. To bridge this gap, we introduce a zero-shot prompting framework, Foresee and Reflect (FaR), which provides a reasoning structure that encourages LLMs to anticipate future challenges and reason about potential actions. FaR boosts GPT-4's performance from 50% to 71% on T4D, outperforming other prompting methods such as Chain-of-Thought and Self-Ask. Moreover, FaR generalizes to diverse out-of-distribution story structures and scenarios that also require ToM inferences to choose an action, consistently outperforming other methods including few-shot in-context learning.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2310.03051

PDF

https://arxiv.org/pdf/2310.03051

OA Status

green

Cited By

6

Related Works

10

OpenAlex ID

https://openalex.org/W4387427450

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4387427450

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2310.03051

Digital Object Identifier

Title

How FaR Are Large Language Models From Agents with Theory-of-Mind?

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2023

Year of publication

Publication date

2023-10-04

Full publication date if available

Authors

Pei Zhou, Aman Madaan, Srividya Pranavi Potharaju, Aditya Gupta, Kevin R. McKee, Ari Holtzman, Jay Pujara, Xiang Ren, Swaroop Mishra, Aida Nematzadeh, Shyam Upadhyay, Manaal Faruqui

List of authors in order

Landing page

https://arxiv.org/abs/2310.03051

Publisher landing page

PDF URL

https://arxiv.org/pdf/2310.03051

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2310.03051

Direct OA link when available

Concepts

Action (physics), Context (archaeology), Shot (pellet), Inference, Computer science, Cognitive science, Artificial intelligence, Psychology, Epistemology, History, Archaeology, Organic chemistry, Chemistry, Quantum mechanics, Physics, Philosophy

Top concepts (fields/topics) attached by OpenAlex

Cited by

6

Total citation count in OpenAlex

Citations by year (recent)

2025: 2, 2024: 3, 2023: 1

Per-year citation counts (last 5 years)

Related works (count)

10

Other works algorithmically related by OpenAlex

Full payload