Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
Oracle-RLAIF: An Improved Fine-Tuning Framework for Multi-modal Video Models using Reinforcement Learning from Ranking FeedbackMicrosoft · arXiv (Cornell University)Oct 2, 2025ACON: Optimizing Context Compression for Long-horizon LLM AgentsMicrosoft · arXiv (Cornell University)150 citationsOct 1, 2025The impact of advanced AI systems on democracyAnthropic, Google DeepMind, OpenAI · Nature Human Behaviour65 citationsOct 1, 2025Characterizing Mobile SoC for Accelerating Heterogeneous LLM InferenceSenseTime · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)43 citationsOct 1, 2025METIS: Fast Quality-Aware RAG Systems with Configuration AdaptationMicrosoft · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)38 citationsOct 1, 2025Eliciting Secret Knowledge from Language ModelsAnthropic · arXiv (Cornell University)34 citationsOct 1, 2025Rethinking Thinking Tokens: LLMs as Improvement OperatorsAnthropic · arXiv (Cornell University)25 citationsOct 1, 2025LithOS: An Operating System for Efficient Machine Learning on GPUsMeta · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)24 citationsOct 1, 2025A Practitioner's Guide to Multi-turn Agentic Reinforcement LearningNVIDIA · arXiv (Cornell University)21 citationsOct 1, 2025Improving Code Localization with Repository MemoryMicrosoft · arXiv (Cornell University)18 citationsOct 1, 2025ThinKV: Thought-Adaptive KV Cache Compression for Efficient Reasoning ModelsNVIDIA · arXiv (Cornell University)17 citationsOct 1, 2025CHERIoT RTOS: An OS for Fine-Grained Memory-Safe Compartments on Low-Cost Embedded DevicesMicrosoft · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)15 citationsOct 1, 2025IC-Cache: Efficient Large Language Model Serving via In-context CachingAlphabet · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)15 citationsOct 1, 2025Making, not Taking, the Best of NCohere · arXiv (Cornell University)12 citationsOct 1, 2025Terabyte-Scale Analytics in the Blink of an EyeMicrosoft · Proceedings of the VLDB Endowment11 citationsOct 1, 2025Rethinking Reward Models for Multi-Domain Test-Time ScalingMicrosoft · arXiv (Cornell University)11 citationsOct 1, 2025TrainVerify: Equivalence-Based Verification for Distributed LLM TrainingMicrosoft · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)9 citationsOct 1, 2025Differentiable Logic Cellular Automata: From Game of Life to Pattern GenerationAlphabet · ALIFE8 citationsOct 1, 2025Oasis: Pooling PCIe Devices Over CXL to Boost UtilizationMicrosoft · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)8 citationsOct 1, 2025Mercury: Unlocking Multi-GPU Operator Optimization for LLMs via Remote Memory SchedulingMeta, OpenAI · ACM SIGOPS Symposium on Operating Systems Principles (SOSP)8 citationsOct 1, 2025The role of graph topology in the performance of biomedical Knowledge Graph Completion modelsGraphcore · Bioinformatics7 citationsOct 1, 2025Data Selection for Fine-tuning Vision Language Models via Cross Modal Alignment TrajectoriesAlphabet · arXiv (Cornell University)7 citationsOct 1, 2025Simultaneous Multi-objective Alignment Across Verifiable and Non-verifiable RewardsNVIDIA · arXiv (Cornell University)7 citationsOct 1, 2025Is Model Editing Built on Sand? Revealing Its Illusory Success and Fragile FoundationAlphabet, Google DeepMind · arXiv (Cornell University)6 citationsOct 1, 2025HARPA: A Testability-Driven, Literature-Grounded Framework for Research IdeationAllen Institute for AI · arXiv (Cornell University)6 citationsOct 1, 2025