Science ❯ Computer Science ❯ Artificial Intelligence

Model Training

Reinforcement Learning Deep Learning Data Poisoning Synthetic Data Generation Conversational Agents Data Challenges Reasoning Models Parameter-efficient Fine-tuning Learning Behaviors Data Annotation Ethical Considerations Simulation Reasoning Traces Inference Optimization Counterfactual Learning

Anthropic-Led Study Finds About 250 Poisoned Documents Can Backdoor LLMs Regardless of Size

Routine safety training can largely neutralize such simple backdoors.