Reflection Steering: Disentangling Reflection from Reasoning in Activation Space for Token-Efficient Inference
cs.LG, cs.CL
Submitted: 2026-08-26
Updated: 2026-08-26
Terminology
Sources
- SEAL: Steerable Reasoning Calibration of Large Language Models for Free
- Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Scaling Reasoning, Losing Control: Evaluating Instruction Following in Large Reasoning Models
- Token-Budget-Aware LLM Reasoning
- O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning
- s1: Simple test-time scaling
- Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters
- Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models
- Steering Language Models With Activation Engineering
- TokenSkip: Controllable Chain-of-Thought Compression in LLMs
- Chain of Draft: Thinking Faster by Writing Less
- ReflCtrl: Controlling LLM Reflection Efficiently via Representation Engineering
- AdaptThink: Reasoning Models Can Learn When to Think
- Understanding and Steering the Cognitive Behaviors of Reasoning Models at Test-Time
- Representation Engineering: A Top-Down Approach to AI Transparency
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks