AI
Projector Is All You Train
Researchers from the University of California, Berkeley, have published a paper proposing a new approach to training multimodal large language models....
AI
AI
One Success Isn't Reliability: Thinkingbox, a Sandbox and Benchmark for Agents in Stateful Business Workflows
Researchers have created a sandbox called Thinkingbox to test the reliability of artificial agents in complex business workflows. The sandbox allows f...
AI
PersonalBench: Measuring the Authorship Gap in LLM Personalization
Researchers have introduced PersonalBench, a benchmark for evaluating personalized text generation in large language models (LLMs). The benchmark asse...
AI
FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving
Researchers have developed FlashPrefill V2, an improved version of their previous work on long-context modeling for Large Language Models. The new sys...
AI
SWE-bench Science: Can Coding Agents Resolve Engineering Tasks in Science?
Researchers have created SWE-bench Science, a benchmark to evaluate coding agents' ability to resolve engineering tasks in science. The benchmark cons...
AI
Stopping and Routing LLM Judge Panels
Researchers from Bin Zhu, Yi Xie, and Yanghui Rao have proposed a method to design judge panels for Large Language Models (LLMs). The approach, called...
AI
Learning how to Forget: Fine-tuning for Long-Context Sparse Attention
Researchers from AWS Labs have developed a new method for fine-tuning transformer language models with sparse attention. Their approach allows models ...
AI
Dynamic Gated Cross-Modal Fusion with Sarcastic-aware Contrastive Regularization for Multimodal Sarcasm Detection
Researchers have developed a new framework for detecting sarcasm in text and images. The framework uses a combination of techniques to identify incons...
AI
Natural Language Code Retrieval for 1C:Enterprise: An Open Benchmark and Efficient Bi-Encoder
Researchers have developed an open benchmark for natural language code retrieval in the 1C:Enterprise ecosystem. This system combines Russian syntax w...
AI
Robust Incomplete Multimodal Sentiment Analysis via Iterative Proxy Correction
Researchers have developed a new approach to multimodal sentiment analysis that can handle incomplete or corrupted inputs. Their method, called iterat...
AI
HealMed: Multilingual Evaluation of Large Language Models in Medicine
Researchers have developed a benchmark called HealMed to evaluate the performance of large language models in medicine across multiple languages. The ...
AI
Auditing Cross-Lingual Fairness in Language Model Watermarking
Researchers propose a framework for evaluating the fairness of language models across different languages. Current evaluation methods focus on English...