About
Research & Development
Updating LLM Knowledge: A Two-Step Approach to Overcome Factual Conflicts
TLDR: A new "unlearn-then-learn" strategy, using the IA3 fine-tuning technique and circuit localization, allows for precise and localized knowledge editing in compact LLMs. This...
Research & Development
Advancing Generative Models with Second-Order MeanFlow
TLDR: A new research paper introduces Second-Order MeanFlow, an extension to generative models that incorporates average acceleration fields. This framework is proven feasible for...
Research & Development
Pref-GUIDE: Smarter AI Training Through Structured Human Preferences
TLDR: Pref-GUIDE is a new framework for training AI agents using real-time human feedback. It converts noisy scalar feedback into consistent pairwise preferences (Pref-GUIDE...
Research & Development
The Unseen Costs of ASR Bias: Why Misrecognition is More Than a Technical Glitch
TLDR: This paper argues that biases in Automatic Speech Recognition (ASR) systems, which systematically misrecognize certain speech varieties, constitute a form of disrespect and...
Research & Development
Unveiling LLM Limitations with Dynamic Code Benchmarking
TLDR: CODE2BENCH is a new framework for creating dynamic, contamination-resistant benchmarks from real-world GitHub code to evaluate Large Language Models (LLMs) on coding tasks....
Research & Development
Crafting AI: How Model Bending in ComfyUI Empowers Artists
TLDR: A research paper introduces a ComfyUI plugin that allows artists to directly manipulate the internal components of large text-to-image diffusion models. This "model...