ข่าว
Whitening Reveals Cluster Commitment as the Geometric Separator of Hallucination Types
arXiv:2603.07755v1 Announce Type: new Abstract: A geometric hallucination taxonomy distinguishes three failure types — center-drift...
Which Questions Improve Learning the Most? Utility Estimation of Questions with LM-based Simulations
arXiv:2502.17383v2 Announce Type: replace Abstract: Asking good questions is critical for comprehension and learning, yet...
Which LLM should you use? Token Monster automatically combines multiple models and tools for you
This architecture lets Token Monster tap into a range of models from different providers without...
When your LLM calls the cops: Claude 4’s whistle-blow and the new agentic AI risk stack
Claude 4’s “whistle-blow” surprise shows why agentic AI risk lives in prompts and tool access...
When your AI browser becomes your enemy: The Comet security disaster
Remember when browsers were simple? You clicked a link, a page loaded, maybe you filled...
When the solution defines the problem
How AI is reversing traditional problem-solving — and uncovering opportunities we never knew existed Continue reading on...
When Shared Knowledge Hurts: Spectral Over-Accumulation in Model Merging
arXiv:2602.05536v1 Announce Type: cross Abstract: Model merging combines multiple fine-tuned models into a single model...
When retrieval outperforms generation: Dense evidence retrieval for scalable fake news detection
arXiv:2511.04643v1 Announce Type: new Abstract: The proliferation of misinformation necessitates robust yet computationally efficient fact...
When Long Helps Short: How Context Length in Supervised Fine-tuning Affects Behavior of Large Language Models
arXiv:2509.18762v1 Announce Type: new Abstract: Large language models (LLMs) have achieved impressive performance across natural...
When Facts Change: Probing LLMs on Evolving Knowledge with evolveQA
arXiv:2510.19172v1 Announce Type: new Abstract: LLMs often fail to handle temporal knowledge conflicts–contradictions arising when...
When Does Personality Composition Matter for Multi-Agent LLM Teams?
arXiv:2606.27443v2 Announce Type: replace-cross Abstract: Personality prompting shapes how large language models communicate, yet whether...
When Annotators Agree but Labels Disagree: The Projection Problem in Stance Detection
arXiv:2603.24231v2 Announce Type: replace Abstract: Stance detection is nearly always formulated as classifying text into...
