Steering Where to Listen: Instruction-Based Activation Steering Redirects Temporal Attention in Large Audio-Language Models
cs.SD, cs.AI, eess.AS
Submitted: 2026-06-09
Updated: 2026-06-09
Terminology
Sources
- Qwen2-Audio Technical Report
- Audio Flamingo 3: Advancing Audio Intelligence with Fully Open Large Audio Language Models
- Gemma 3 Technical Report
- Step-Audio 2 Technical Report
- Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers
- SpeechGPT: Empowering Large Language Models with Intrinsic Cross-Modal Conversational Abilities
- Moshi: a speech-text foundation model for real-time dialogue
- LLaMA-Omni: Seamless Speech Interaction with Large Language Models
- Representation Engineering: A Top-Down Approach to AI Transparency
- Steering Language Models With Activation Engineering
- Silence is Golden: Mitigating Hallucinations in Large Audio-Language Models via Layer-Weighted Vector Steering
- Listening Between the Frames: Bridging Temporal Gaps in Large Audio-Language Models
- Not in Sync: Unveiling Temporal Bias in Audio Chat Models
Related papers
- Speech-DRAME: A Framework for Human-Aligned Benchmarks in Speech Role-Play
- MuLaCover: Flexible Cover Song Generation via Symbolic Tonal Control
- Optimising Neural Speech Codecs for 300bps Communication using Reinforcement Learning
- Few-Shot Open-Set Audio Classification via Transductive Prototype Refinement and Class Logit Enhancement
- Spectral Masking and Interpolation Attack (SMIA): A Black-box Adversarial Attack against Voice Authentication and Anti-Spoofing Systems
- Remix the Timbre: Diffusion-Based Style Transfer Across Polyphonic Stems