TACT: Mitigating Overthinking and Overacting in Coding Agents via Activation Steering
cs.AI
Submitted: 2026-05-07
Updated: 2026-11-10
Terminology
Sources
- The Danger of Overthinking: Examining the Reasoning-Action Dilemma in Agentic Tasks
- SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?
- A Comprehensive Survey on Benchmarks and Solutions in Software Engineering of LLM-Empowered Agentic System
- Kimi k1.5: Scaling Reinforcement Learning with LLMs
- Programming Refusal with Conditional Activation Steering
- The Assistant Axis: Situating and Stabilizing the Default Persona of Language Models
- Terminal-Bench: Benchmarking Agents on Hard, Realistic Tasks in Command Line Interfaces
- s1: Simple test-time scaling
- Agent Drift: Quantifying Behavioral Degradation in Multi-Agent LLM Systems Over Extended Interactions
- A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models
- Conversation for Non-verifiable Learning: Self-Evolving LLMs through Meta-Evaluation
- Meta-Reasoner: Dynamic Guidance for Optimized Inference-time Reasoning in Large Language Models
- What-If Analysis of Large Language Models: Explore the Game World Using Proactive Thinking
- Steering Language Models With Activation Engineering
- Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement
- Position: Agent Should Invoke External Tools ONLY When Epistemically Necessary
- Why Reasoning Fails to Plan: A Planning-Centric Analysis of Long-Horizon Decision Making in LLM Agents
- Live-SWE-agent: Can Software Engineering Agents Self-Evolve on the Fly?
- Claw-Eval: Towards Trustworthy Evaluation of Autonomous Agents
- Survey on Evaluation of LLM-based Agents
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents