Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
CUSP: Cutoff-conditioned Unseen Scientific ProgressAllen Institute for AI, Sakana AI · arXiv (Cornell University)Aug 16Ask-E: An Environment for Calibrated Question GenerationAllen Institute for AI · arXiv (Cornell University)Aug 7MameLoshnLM: Yiddish Language Model and Evaluation BenchmarkAllen Institute for AI · arXiv (Cornell University)Aug 6A Dual Evaluation for Music TranscriptionAllen Institute for AI · arXiv (Cornell University)Aug 5Foundations of Stochastic Lexical Calculus: Semantic Descent and Random Dynamics on Probability SimplicesAllen Institute for AI · arXiv (Cornell University)Jul 25Identification and Honest Recovery from Semantic Observation Kernels: Operator Error, Coarsening, and StabilityAllen Institute for AI · arXiv (Cornell University)Jul 25Co-Harness: Co-Evolving Harnesses and Model Weights for LLM AgentsAllen Institute for AI · arXiv (Cornell University)Jul 17Pretraining Data Can Be Poisoned through Computational PropagandaAllen Institute for AI · arXiv (Cornell University)Jul 16Knowledge- and Gradient-Guided Reinforcement Learning for Parametrized Action Markov Decision ProcessesAllen Institute for AI · arXiv (Cornell University)Jul 14Rethinking the Evaluation of Harness Evolution for AgentsAllen Institute for AI · arXiv (Cornell University)Jul 14Perspective on Self-Supervised Transformers for Scalable Data Analysis, Retrievals, and Discovery in GeosciencesAllen Institute for AI · Artificial Intelligence for the Earth SystemsJul 13GLANCED-IO: Taming I/O Optimization for Deep Learning at ScaleAllen Institute for AI · ACM International Symposium on High-Performance Parallel and Distributed Computing (HPDC)Jul 11AppWorld-UL: Benchmarking Diverse Agent-User Interactions for Tool-UseAllen Institute for AI · arXiv (Cornell University)Jul 10LEGATO 2: Toward Multimodal Sheet Music Recognition and UnderstandingAllen Institute for AI · arXiv (Cornell University)Jul 7Do Vision-Language-Action Models Mean What They Say? On the Role of Faithfulness in Embodied ReasoningAllen Institute for AI, NVIDIA · arXiv (Cornell University)Jul 6MANCE: Manifold Aware Concept ErasureAllen Institute for AI · arXiv (Cornell University)Jul 4Evidence-Informed LLM Beliefs for Continual Scientific DiscoveryAllen Institute for AI · arXiv (Cornell University)Jun 28Agentic Abstention: Do Agents Know When to Stop Instead of Act?Allen Institute for AI · arXiv (Cornell University)Jun 27A Human-Centric Framework for Data Attribution in Large Language ModelsAllen Institute for AI · ACM Conference on Fairness, Accountability, and Transparency (FAccT)Jun 23Tmax: A simple recipe for terminal agentsAllen Institute for AI · arXiv (Cornell University)Jun 22Comparing Transformers and Hybrid Models at the Token LevelAllen Institute for AI · arXiv (Cornell University)Jun 18Efficiently Representing Algorithms With Chain-of-Thought TransformersAllen Institute for AI · arXiv (Cornell University)Jun 18ReFinE: Streamlining UI Mockup Iteration with Research FindingsAllen Institute for AI · Designing Interactive Systems ConferenceJun 12Operadic consistency: a label-free signal for compositional reasoning failures in LLMsAllen Institute for AI · arXiv (Cornell University)Jun 11Which Models Are Our Models Built On? Auditing Invisible Dependencies in Modern LLMsAllen Institute for AI · arXiv (Cornell University)Jun 10