UpSafe C: Upcycling for Controllable Safety in Large Language Models
cs.AI, cs.CR, cs.LG
Submitted: 2025-10-02
Updated: 2026-11-04
Code: https://github.com/huggingface/peft
Terminology
Sources
- GPT-4 Technical Report
- Evaluating Large Language Models Trained on Code
- Attacks, Defenses and Evaluations for LLM Conversation Safety: A Survey
- MoGU: A Framework for Enhancing Safety of Open-Sourced LLMs While Preserving Their Usability
- KTO: Model Alignment as Prospect Theoretic Optimization
- Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned
- MART: Improving LLM Safety with Multi-round Automatic Red-Teaming
- The Llama 3 Herd of Models
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Upcycling Large Language Models into Mixture of Experts
- Measuring Massive Multitask Language Understanding
- Curiosity-driven Red-teaming for Large Language Models
- Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
- Mixtral of Experts
- SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities
- How Large Language Models Encode Context Knowledge? A Layer-Wise Probing Study
- Scaling Laws for Neural Language Models
- Sparse Upcycling: Training Mixture-of-Experts from Dense Checkpoints
- Safety Layers in Aligned Large Language Models: The Key to LLM Security
- ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversation
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