majority voting
A simple ensemble method where multiple predictions from different models are combined by selecting the most frequent output, ensuring robustness and potentially improving accuracy by aggregating diverse model perspectives.
- Class conditional conformal prediction for multiple inputs by p-value aggregation
- Cost-aware LLM-based Online Dataset Annotation
- Debate or Vote: Which Yields Better Decisions in Multi-Agent Large Language Models?
- First SFT, Second RL, Third UPT: Continual Improving Multi-Modal LLM Reasoning via Unsupervised Post-Training
- Let Me Think! A Long Chain of Thought Can Be Worth Exponentially Many Short Ones
- Median Selection with Noisy and Structural Information
- TTRL: Test-Time Reinforcement Learning
- The Overthinker's DIET: Cutting Token Calories with DIfficulty-AwarE Training
- Value-Guided Search for Efficient Chain-of-Thought Reasoning