REVIEW 7 cited by
Explore Theory of Mind: Program-guided adversarial data generation for theory of mind reasoning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Do large language models (LLMs) have theory of mind? A plethora of papers and benchmarks have been introduced to evaluate if current models have been able to develop this key ability of social intelligence. However, all rely on limited datasets with simple patterns that can potentially lead to problematic blind spots in evaluation and an overestimation of model capabilities. We introduce ExploreToM, the first framework to allow large-scale generation of diverse and challenging theory of mind data for robust training and evaluation. Our approach leverages an A* search over a custom domain-specific language to produce complex story structures and novel, diverse, yet plausible scenarios to stress test the limits of LLMs. Our evaluation reveals that state-of-the-art LLMs, such as Llama-3.1-70B and GPT-4o, show accuracies as low as 0% and 9% on ExploreToM-generated data, highlighting the need for more robust theory of mind evaluation. As our generations are a conceptual superset of prior work, fine-tuning on our data yields a 27-point accuracy improvement on the classic ToMi benchmark (Le et al., 2019). ExploreToM also enables uncovering underlying skills and factors missing for models to show theory of mind, such as unreliable state tracking or data imbalances, which may contribute to models' poor performance on benchmarks.
Forward citations
Cited by 7 Pith papers
-
TimeHC-RL: Temporal-aware Hierarchical Cognitive Reinforcement Learning for Enhancing LLMs' Social Intelligence
A temporal-aware hierarchical reinforcement learning method improves a 7B LLM's social reasoning enough to rival DeepSeek-R1 and OpenAI-O3 on in-domain theory-of-mind benchmarks.
-
S-MARC: Causal Streaming Reasoning for Full-Duplex Conversational Behavior Modeling
A streaming causal model predicts per-second two-level speech acts and rationale explanations, trained on 120 hours of LLM-synthesized duplex dialogue.
-
Small LLMs Do Not Learn a Generalizable Theory of Mind via Reinforcement Learning
Reinforcement learning with verifiable rewards makes a small LLM overfit theory-of-mind benchmarks, not acquire a generalizable theory of mind.
-
Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework
A framework paper that adapts AI safety case methodology to the specific threat of manipulation attacks by internally deployed misaligned AI.
-
The Decrypto Benchmark for Multi-Agent Reasoning and Theory of Mind
A new interactive language-game benchmark shows LLMs lag behind simple word-embedding baselines and that newer reasoning models regress on theory-of-mind tasks.
-
Data Swarms: Optimizable Generation of Synthetic Evaluation Data
Data Swarms uses particle swarm optimization over data-generator LLM weights to produce synthetic evaluation data that scores higher on five quantitative evaluation objectives than eight baselines.
-
Embodied AI Agents: Modeling the World
Embodied AI agents should be built around physical world models plus a mental world model of the user, with virtual, wearable, and robotic agents sharing this core.
Discussion (0). Sign in to comment.