Research
A paper is listed when one of its authors is affiliated with the lab. Data: OpenAlex.
jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual RetrievalJina AI · Workshop on Multilingual Representation Learning (MRL)35 citationsJan 1, 2025SSR: Enhancing Depth Perception in Vision-Language Models via Rationale-Guided Spatial ReasoningAlibaba · neural information processing systems35 citationsJan 1, 2025ResearchCodeBench: Benchmarking LLMs on Implementing Novel Machine Learning Research CodeOpenAI · neural information processing systems35 citationsJan 1, 2025The Curse of Multi-Modalities: Evaluating Hallucinations of Large Multimodal Models across Language, Visual, and AudioAlibaba · neural information processing systems35 citationsJan 1, 2025Self-Improving Embodied Foundation ModelsAlphabet, Google DeepMind · neural information processing systems34 citationsJan 1, 2025Contextual Integrity in LLMs via Reasoning and Reinforcement LearningMicrosoft · neural information processing systems34 citationsJan 1, 2025TokenSelect: Efficient Long-Context Inference and Length Extrapolation for LLMs via Dynamic Token-Level KV Cache SelectionAlibaba · Conference on Empirical Methods in Natural Language Processing (EMNLP)34 citationsJan 1, 2025Ensembling Large Language Models with Process Reward-Guided Tree Search for Better Complex ReasoningMicrosoft · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)34 citationsJan 1, 2025ALE-Bench: A Benchmark for Long-Horizon Objective-Driven Algorithm EngineeringSakana AI · neural information processing systems33 citationsJan 1, 2025Model Merging in Pre-training of Large Language ModelsAlphabet · neural information processing systems33 citationsJan 1, 2025SciRIFF: A Resource to Enhance Language Model Instruction-Following over Scientific LiteratureAllen Institute for AI · Conference on Empirical Methods in Natural Language Processing (EMNLP)33 citationsJan 1, 2025SceneWeaver: All-in-One 3D Scene Synthesis with an Extensible and Self-Reflective AgentAlibaba · neural information processing systems33 citationsJan 1, 2025Defending against Indirect Prompt Injection by Instruction DetectionMicrosoft · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)32 citationsJan 1, 2025Hybrid Latent Reasoning via Reinforcement LearningAlphabet, Meta, SenseTime · neural information processing systems32 citationsJan 1, 2025ControlSpeech: Towards Simultaneous and Independent Zero-shot Speaker Cloning and Zero-shot Language Style ControlAlibaba · Proceedings of the conference - Association for Computational Linguistics. Meeting32 citationsJan 1, 2025The Hawthorne Effect in Reasoning Models: Evaluating and Steering Test AwarenessMicrosoft · neural information processing systems31 citationsJan 1, 2025Provable Scaling Laws for the Test-Time Compute of Large Language ModelsAlibaba · neural information processing systems31 citationsJan 1, 2025Look Before You Leap: A GUI-Critic-R1 Model for Pre-Operative Error Diagnosis in GUI AutomationAlibaba · neural information processing systems31 citationsJan 1, 20254KAgent: Agentic Any Image to 4K Super-ResolutionAlphabet, Google DeepMind · neural information processing systems31 citationsJan 1, 2025Parallel Scaling Law for Language ModelsAlibaba · neural information processing systems31 citationsJan 1, 2025Prismatic Synthesis: Gradient-based Data Diversification Boosts Generalization in LLM ReasoningNVIDIA · neural information processing systems31 citationsJan 1, 2025Assessing Dialect Fairness and Robustness of Large Language Models in Reasoning TasksAllen Institute for AI, Microsoft · Proceedings of the conference - Association for Computational Linguistics. Meeting31 citationsJan 1, 2025K-Level Reasoning: Establishing Higher Order Beliefs in Large Language Models for Strategic ReasoningMicrosoft · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT)31 citationsJan 1, 2025UniGen: Enhanced Training & Test-Time Strategies for Unified Multimodal Understanding and GenerationAlphabet · neural information processing systems30 citationsJan 1, 2025Power Lines: Scaling laws for weight decay and batch size in LLM pre-trainingCerebras Systems · neural information processing systems30 citationsJan 1, 2025