Refusal steering works on MoE LLMs; expert-aware variants succeed with single-expert outputs and refusal signals differ from routing patterns.
A closer look at the limitations of instruction tuning.arXiv preprint arXiv:2402.05119
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
representative citing papers
Distilling safe refusal behavior from OpenAI o1-mini into Llama-3, Gemma-2, and Qwen3 models via response-based LoRA on multilingual jailbreak data increases jailbreak success rates on MultiJail by up to 16.6 points.
COMPACT synthesizes compositional visual instruction data to reduce VIT training data by 90% while achieving 100.2% of full performance across eight multimodal benchmarks.
A 3B model with few-shot prompting reaches 79.7% of GPT-5 tool-use performance while a hypernetwork adaptation adds zero measurable benefit across four benchmarks.
Embedding-based QLoRA fine-tuning of causal LLMs matches BERT on single-label patent classification with 10–30x fewer trainable parameters, while instruction-tuning wins on multi-label classification only with ≥100M trainable parameters.
citing papers explorer
-
Expert-Aware Refusal Steering
Refusal steering works on MoE LLMs; expert-aware variants succeed with single-expert outputs and refusal signals differ from routing patterns.
-
Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety
Distilling safe refusal behavior from OpenAI o1-mini into Llama-3, Gemma-2, and Qwen3 models via response-based LoRA on multilingual jailbreak data increases jailbreak success rates on MultiJail by up to 16.6 points.
-
Visual Compositional Tuning
COMPACT synthesizes compositional visual instruction data to reduce VIT training data by 90% while achieving 100.2% of full performance across eight multimodal benchmarks.
-
Meta-Tool: Efficient Few-Shot Tool Adaptation for Small Language Models
A 3B model with few-shot prompting reaches 79.7% of GPT-5 tool-use performance while a hypernetwork adaptation adds zero measurable benefit across four benchmarks.
-
Fine-Tuning Causal LLMs for Text Classification: Embedding-Based vs. Instruction-Based Approaches
Embedding-based QLoRA fine-tuning of causal LLMs matches BERT on single-label patent classification with 10–30x fewer trainable parameters, while instruction-tuning wins on multi-label classification only with ≥100M trainable parameters.