When Keywords Drop but Classifiers Hold: Soft Refusals under KV Cache Compression
cs.LG, cs.CL
Submitted: 2026-09-16
Updated: 2026-09-16
Comments: 8 pages, 5 figures
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- The Llama 3 Herd of Models
- Measuring Massive Multitask Language Understanding
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
- AnchorKV: Safety-Aware KV Cache Compression via Soft Penalty with a Refusal Anchor
- Alignment Collapse Under KV Cache Quantization: Diagnosis and Mitigation
- Qwen2 Technical Report
- Universal and Transferable Adversarial Attacks on Aligned Language Models
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: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks