Neologism Inoculation Prompting
cs.AI, cs.CL
Submitted: 2026-11-03
Updated: 2026-11-03
Terminology
Sources
- Recontextualization Mitigates Specification Gaming without Modifying the Specification
- Skill Neologisms: Towards Skill-based Continual Learning
- Gradient Routing: Masking Gradients to Localize Computation in Neural Networks
- Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
- Don't Inoculate Everything: Stratified Inoculation Prompting Narrows Backdoor Triggers and Preserves Desired Traits
- Neologism Learning for Controllability and Self-Verbalization
- CTRL: A Conditional Transformer Language Model for Controllable Generation
- Mark, Don't Erase: Token Inoculation for Dual-Use Knowledge in LLMs
- Natural Emergent Misalignment from Reward Hacking in Production RL
- Inoculation Midtraining with Learned Neologisms
- Neologism Learning as a Parameter-Efficient Alternative to Fine-Tuning for Model Steering
- Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors
- Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time
- School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs
- Model Organisms for Emergent Misalignment
- Gradient Routing Localizes and Removes Unintended Behaviors in RL
- Inoculation Prompting: Instructing LLMs to misbehave at train-time improves test-time alignment
- Universal and Transferable Adversarial Attacks on Aligned Language Models
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