6 papers
Cross-LLM Consistency in Inference: Evidence from Shared Interactions
Siyu Lou, Yao Yan, Yuntian Chen +1
Large language models (LLMs) differ in architecture, training data, and optimization procedures, yet they may still develop similar internal inference patterns. In this paper, we e…
Towards a Science of Scaling Agent Systems
Yubin Kim, Ken Gu, Chanwoo Park +17
Agents, language model-based systems capable of reasoning, planning, and acting are widely adopted in real-world tasks, yet how their performance changes as these systems scale acr…
Post-Routing Arithmetic in Llama-3: Last-Token Result Writing and Rotation-Structured Digit Directions
Yao Yan
We study three-digit addition in Meta-Llama-3-8B (base) under a one-token readout to characterize how arithmetic answers are finalized after cross-token routing becomes causally ir…
Exploring Human-AI Conceptual Alignment through the Prism of Chess
Semyon Lomasov, Judah Goldfeder, Mehmet Hamza Erol +5
Do AI systems truly understand human concepts or merely mimic surface patterns? We investigate this through chess, where human creativity meets precise strategic concepts. Analyzin…
ScaleCUA: Scaling Open-Source Computer Use Agents with Cross-Platform Data
Zhaoyang Liu, Jingjing Xie, Zichen Ding +27
Vision-Language Models (VLMs) have enabled computer use agents (CUAs) that operate GUIs autonomously, showing great potential, yet progress is limited by the lack of large-scale, o…
Addition in Four Movements: Mapping Layer-wise Information Trajectories in LLMs
Yao Yan
Multi-digit addition is a clear probe of the computational power of large language models. To dissect the internal arithmetic processes in LLaMA-3-8B-Instruct, we combine linear pr…