AI.info

Kevin Zhu

Explore Kevin Zhu on AI.info.

• SuperNeuroMAT: An Efficient Matrix-based Simulator for Spiking Neural Networks
• A Few Bad Neurons: Isolating and Surgically Correcting Sycophancy
• AMVICC: A Novel Benchmark for Cross-Modal Failure Mode Profiling for VLMs and IGMs
• CBMAS: Cognitive Behavioral Modeling via Activation Steering
• Emergent Persuasion: Will LLMs Persuade Without Being Prompted?
• Emergent World Beliefs: Exploring Transformers in Stochastic Games
• TorchTraceAP: A New Benchmark Dataset for Detecting Performance Anti-Patterns in Computer Vision Models
• Reasoning Relay: Evaluating Stability and Interchangeability of Large Language Models in Mathematical Reasoning
• Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language Models
• WOLF: Werewolf-based Observations for LLM Deception and Falsehoods
• ASCIIBench: Evaluating Language-Model-Based Understanding of Visually-Oriented Text
• Sumudu Neural Operator for ODEs and PDEs
• Grounding Foundational Vision Models with 3D Human Poses for Robust Action Recognition
• Inference-Time Chain-of-Thought Pruning with Latent Informativeness Signals
• SMAGDi: Socratic Multi Agent Interaction Graph Distillation for Efficient High Accuracy Reasoning
• DynaStride: Dynamic Stride Windowing with MMCoT for Instructional Multi-Scene Captioning
• DuoLens: A Framework for Robust Detection of Machine-Generated Multilingual Text and Code
• AgentChangeBench: A Multi-Dimensional Evaluation Framework for Goal-Shift Robustness in Conversational AI
• Interpreting the Latent Structure of Operator Precedence in Language Models