From the 1 of 5 linked papers with an AI index.
5 papers
A Shared Subcircuit Lets LLMs Count Down Across Tasks
Jacob Dunefsky, Wes Gurnee, Emmanuel Ameisen
The paper identifies a "countdown subcircuit" in Llama-3.1-70B-Instruct that compares the current token position to a target length, enabling the model to track remaining tokens fo…
Combatting Gerrymandering with Ranked Choice Voting: An Experimental Analysis of Multi-member Districts in the United States
Nikhil Garg, Wes Gurnee, David Rothschild +1
Every representative democracy must specify a mechanism under which voters choose their representatives. The most common mechanism in the United States -- Winner takes all single-m…
When Models Manipulate Manifolds: The Geometry of a Counting Task
Wes Gurnee, Emmanuel Ameisen, Isaac Kauvar +4
Language models can perceive visual properties of text despite receiving only sequences of tokens-we mechanistically investigate how Claude 3.5 Haiku accomplishes one such task: li…
The Remarkable Robustness of LLMs: Stages of Inference?
Vedang Lad, Jin Hwa Lee, Wes Gurnee +1
We investigate the robustness of Large Language Models (LLMs) to structural interventions by deleting and swapping adjacent layers during inference. Surprisingly, models retain 72-…
Not All Language Model Features Are One-Dimensionally Linear
Joshua Engels, Eric J. Michaud, Isaac Liao +2
Recent work has proposed that language models perform computation by manipulating one-dimensional representations of concepts ("features") in activation space. In contrast, we expl…