7 papers
Causal and Interpretable Structures in LLM Compositional Tasks
Gurbir Arora, Toni J. B. Liu, Jiajun Bao +2
Large language models are able to solve tasks whose answers depend on not only individual input tokens, but also on relations among them. How is such relational information represe…
The Geometry of Ignorance: LLMs Know When to Temper Bayesian Priors
Toni J. B. Liu, Jiajun Bao, Yizhou Liu +4
What does a language model predict when it has few clues? The answer lurks in its unembedding geometry: a single direction of the unembedding matrix encodes the unigram distributio…
A Graph Signal Processing Perspective on Numerical Sequence Representations in LLM In-Context Learning
Jiajun Bao, Zihao Qi, Toni J. B. Liu +4
Pretrained large language models (LLMs) have demonstrated in-context learning (ICL) capabilities for numerical inference over sequences serialized as text. Prior work has identifie…
What LLM Forecasters Know but Don't Say: Probing Internal Representations for Calibration and Faithfulness
Raphaël Sarfati, Pratyush Ranjan Tiwari, Siddharth Boppana +3
Large language models fine-tuned for forecasting can be accurate yet poorly calibrated, and their chain-of-thought (CoT) reasoning may not faithfully reflect the evidence behind a…
Jacobian Scopes: token-level causal attributions in LLMs
Toni J. B. Liu, Baran Zadeoğlu, Nicolas Boullé +3
Large language models (LLMs) make next-token predictions based on clues present in their context, such as semantic descriptions and in-context examples. Yet, elucidating which prio…
Text-Trained LLMs Can Zero-Shot Extrapolate PDE Dynamics, Revealing a Three-Stage In-Context Learning Mechanism
Jiajun Bao, Nicolas Boullé, Toni J. B. Liu +2
Large language models (LLMs) have demonstrated emergent in-context learning (ICL) capabilities across a range of tasks, including zero-shot time-series forecasting. We show that te…
What's in a prompt? Language models encode literary style in prompt embeddings
Raphaël Sarfati, Haley Moller, Toni J. B. Liu +2
Large language models use high-dimensional latent spaces to encode and process textual information. Much work has investigated how the conceptual content of words translates into g…