5 papers
Annotations Mitigate Post-Training Mode Collapse
Jacob Mitchell Springer, Madhu Advani, Lukas Aichberger +7
Post-training (via supervised fine-tuning) improves instruction-following, but often induces semantic mode collapse by biasing models toward low-entropy fine-tuning data at the exp…
Trained on Tokens, Calibrated on Concepts: The Emergence of Semantic Calibration in LLMs
Preetum Nakkiran, Arwen Bradley, Adam Goliński +3
Large Language Models (LLMs) often lack meaningful confidence estimates for their outputs. While base LLMs are known to exhibit next-token calibration, it remains unclear whether t…
To Infinity and Beyond: Tool-Use Unlocks Length Generalization in State Space Models
Eran Malach, Omid Saremi, Sinead Williamson +5
State Space Models (SSMs) have become the leading alternative to Transformers for sequence modeling. Their primary advantage is efficiency in long-context and long-form generation,…
Trace Length is a Simple Uncertainty Signal in Reasoning Models
Siddartha Devic, Charlotte Peale, Arwen Bradley +3
Uncertainty quantification for LLMs is a key research direction towards addressing hallucination and other issues that limit their reliable deployment. In this work, we show that r…
Local Mechanisms of Compositional Generalization in Conditional Diffusion
Arwen Bradley
Conditional diffusion models appear capable of compositional generalization, i.e., generating convincing samples for out-of-distribution combinations of conditioners, but the mecha…