nowait

Topics in the source

• NOWAIT Reasoning Optimizer
• Overview
• When to Use
• Supported Models
• Quick Start
• 1. Basic Implementation
• 2. Keywords Suppressed
• How It Works

Implements the NOWAIT technique from the paper "Wait, We Don't Need to 'Wait'! Removing Thinking Tokens Improves Reasoning Efficiency" (Wang et al., 2025). NOWAIT is a training-free inference-time intervention that suppresses self-reflection tokens (e.g., "Wait", "Hmm", "Alternatively") during generation, reducing chain-of-thought (CoT) trajectory length by 27-51% without compromising model utility. Deploying R1-style reasoning models with limited compute Reducing inference latency for production systems Optimizing token costs for reasoning tasks Working with verbose CoT outputs that need streamlining | Model Series | Type | Token Reduction | |--------------|------|-----------------| | QwQ-32B | RL-based | 16-31% | | Phi4-Reasoning-Plus | RL-based | 23-28% | | Qwen3-32B | RL-based | 13-16% | | Kimi-VL-A3B | Multimodal | 40-60% | | QvQ-72B-Preview | Multimodal | 20-30% | Important: NOWAIT works best with RL-based models. Distilled models (Qwen3-4B/8B/14B) show degraded performance when reflection tokens are suppressed. See references/keywords.md for the complete list. Core keywords: 1. Initialize Keywords: Identify reflection keywords from empirical analysis 2. Expand to Token Varia

This is an excerpt. Refer to the complete source and its current requirements.