12 papers
Agents of Chaos
Natalie Shapira, Chris Wendler, Avery Yen +35
We report an exploratory red-teaming study of autonomous language-model-powered agents deployed in a live laboratory environment with persistent memory, email accounts, Discord acc…
Internal states before wait modulate reasoning patterns
Dmitrii Troitskii, Koyena Pal, Chris Wendler +2
Prior work has shown that a significant driver of performance in reasoning models is their ability to reason and self-correct. A distinctive marker in these reasoning traces is the…
Discovering Forbidden Topics in Language Models
Can Rager, Chris Wendler, Rohit Gandikota +1
Refusal discovery is the task of identifying the full set of topics that a language model refuses to discuss. We introduce this new problem setting and develop a refusal discovery…
zip2zip: Inference-Time Adaptive Tokenization via Online Compression
Saibo Geng, Nathan Ranchin, Yunzhen yao +4
Tokenization efficiency plays a critical role in the performance and cost of large language models (LLMs), yet most models rely on static tokenizers optimized on general-purpose co…
The Geometry of Self-Verification in a Task-Specific Reasoning Model
Andrew Lee, Lihao Sun, Chris Wendler +2
How do reasoning models verify their own answers? We study this question by training a model using DeepSeek R1's recipe on the CountDown task. We leverage the fact that preference…
Localized Cultural Knowledge is Conserved and Controllable in Large Language Models
Veniamin Veselovsky, Berke Argin, Benedikt Stroebl +5
Just as humans display language patterns influenced by their native tongue when speaking new languages, LLMs often default to English-centric responses even when generating in othe…