Au-M-ol: A Unified Model for Medical Audio and Language Understanding
A research team has introduced Au-M-ol, a multimodal architecture that extends audio processing capabilities to large la…
📂 Research
Latest articles in Research
Small Models Solve the Prompt Ambiguity Problem for Large Language Models
A new arXiv study proposes using Small Language Models (SLMs) to resolve semantic ambiguity in Large Language Model prom…
Fine-Tuning vs. In-Context Learning: Which Is Stronger? New Research From a Formal Language Perspective
A latest arXiv paper systematically compares fine-tuning (FT) and in-context learning (ICL) in large language models fro…
Unveiling the Internal Causes of LLM Jailbreaks: A Study on Layer-wise Feature Vulnerabilities
A new study proposes a three-stage mechanistic analysis pipeline that performs layer-by-layer parsing of internal featur…
Heterogeneous Grouped Mixture-of-Experts Architecture: A New MoE Paradigm Breaking the Uniform Expert Bottleneck
A latest arXiv paper proposes the Mixture of Heterogeneous Grouped Experts architecture, breaking the one-size-fits-all …
ContextWeaver: Building Dependency-Structured Memory for LLM Agents
New research proposes the ContextWeaver framework, which addresses the challenge of LLM agents losing critical structure…
DeepImagine: Training LLMs for Biomedical Reasoning with Counterfactual Imagination
A latest arXiv paper introduces the DeepImagine framework, which teaches large language models biomedical reasoning thro…
LLM Reveals Implicit Language Framing in Obstetric Consultation Notes
A new arXiv study proposes an LLM-based pipeline method to systematically analyze implicit language framing effects in o…
New Study Evaluates Temporal Consistency of Large Language Models in Multi-Turn Conversations
A latest arXiv paper introduces the concept of "Temporal Scope Stability," systematically evaluating large language mode…
SKR Method: Enabling Large Models to Self-Adapt to Tasks Using Intrinsic Knowledge
Researchers propose Self-Knowledge Re-expression (SKR), a method that requires no external data or fine-tuning. By chang…
New Research Introduces Uncertainty Quantification for LLM Function Calling
A latest arXiv paper proposes an uncertainty quantification method for LLM function-calling scenarios, aiming to reduce …
First Chinese Job Skill NER Dataset Released
A research team has released Chinese-SkillSpan, the first named entity recognition dataset for job skills in Chinese rec…