Skip to main content ->
Ai2

Research - Papers

Explore a selection of our published work on a variety of key research challenges in AI.

Filter papers

Answer, Assemble, Ace: Understanding How LMs Answer Multiple Choice Questions

Sarah WiegreffeOyvind TafjordYonatan BelinkovAshish Sabharwal
2025
ICLR

Multiple-choice question answering (MCQA) is a key competence of performant transformer language models that is tested by mainstream benchmarks. However, recent evidence shows that models can have… 

Holistically Evaluating the Environmental Impact of Creating Language Models

Jacob MorrisonClara NaJared FernandezJesse Dodge
2025
ICLR

As the performance of artificial intelligence systems has dramatically increased, so too has the environmental impact of creating these systems. While many model developers release estimates of the… 

On Linear Representations and Pretraining Data Frequency in Language Models

Jack MerulloNoah A. SmithSarah WiegreffeYanai Elazar
2025
ICLR

Pretraining data has a direct impact on the behaviors and quality of language models (LMs), but we only understand the most basic principles of this relationship. While most work focuses on… 

Mechanistic?

Naomi SaphraSarah Wiegreffe
2024
EMNLP • BlackboxNLP Workshop on Analyzing and Interpreting Neural Networks for NLP

The rise of the term “mechanistic interpretability” has accompanied increasing interest in understanding neural models—particularly language models. However, this jargon has also led to a fair… 

SUPER: Evaluating Agents on Setting Up and Executing Tasks from Research Repositories

Ben BoginKejuan YangShashank GuptaTushar Khot
2024
EMNLP

Given that Large Language Models (LLMs) have made significant progress in writing code, can they now be used to autonomously reproduce results from research repositories? Such a capability would be… 

AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents

Harsh TrivediTushar KhotMareike HartmannNiranjan Balasubramanian
2024
ACL

Autonomous agents that address day-to-day digital tasks (e.g., ordering groceries for a household), must not only operate multiple apps (e.g., notes, messaging, shopping app) via APIs, but also… 

PoliFormer: Scaling On-Policy RL with Transformers Results in Masterful Navigators

Kuo-Hao ZengZichen ZhangKiana EhsaniLuca Weihs
2024
CoRL

We present PoliFormer (Policy Transformer), an RGB-only indoor navigation agent trained end-to-end with reinforcement learning at scale that generalizes to the real-world without adaptation despite… 

Universal Visual Decomposer: Long-Horizon Manipulation Made Easy

Zichen ZhangYunshuang LiOsbert BastaniLuca Weihs
2024
IEEE International Conference on Robotics and Automation

Real-world robotic tasks stretch over extended horizons and encompass multiple stages. Learning long-horizon manipulation tasks, however, is a long-standing challenge, and demands decomposing the… 

Mitigating Barriers to Public Social Interaction with Meronymous Communication

Nouran SolimanHyeonsu B. KangMatthew LatzkeDavid R. Karger
2024
CHI

In communities with social hierarchies, fear of judgment can discourage communication. While anonymity may alleviate some social pressure, fully anonymous spaces enable toxic behavior and hide the… 

WildChat: 1M ChatGPT Interaction Logs in the Wild

Wenting ZhaoXiang RenJ. HesselYuntian Deng
2024
ICLR

Chatbots such as GPT-4 and ChatGPT are now serving millions of users. Despite their widespread use, there remains a lack of public datasets showcasing how these tools are used by a population of… 

1-10Next