reasoning abilities
Reasoning abilities refer to the cognitive capabilities of AI models or systems to logically infer relationships, deduce conclusions, and solve problems based on provided information. These capabilities often enhance the model's operational effectiveness in complex scenarios.
- Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
- Constant Bit-size Transformers Are Turing Complete
- Learning to Think: Information-Theoretic Reinforcement Fine-Tuning for LLMs
- MASTER: Enhancing Large Language Model via Multi-Agent Simulated Teaching
- NaDRO: Leveraging Dual-Reward Strategies for LLMs Training on Noisy Data
- Safe + Safe = Unsafe? Exploring How Safe Images Can Be Exploited to Jailbreak Large Vision-Language Models
- Scientists' First Exam: Probing Cognitive Abilities of MLLM via Perception, Understanding, and Reasoning
- SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning
- Strassen Attention, Split VC Dimension and Compositionality in Transformers
- Towards Reliable LLM-based Robots Planning via Combined Uncertainty Estimation
- VITRIX-CLIPIN: Enhancing Fine-Grained Visual Understanding in CLIP via Instruction-Editing Data and Long Captions