A Definition of Catastrophic Performance
cs.AI, cs.LG
Submitted: 2026-11-04
Updated: 2026-11-06
Terminology
Sources
- Mitigating Goal Misgeneralization via Minimax Regret
- Low Impact Artificial Intelligences
- Fragility of Value under Imperfect Alignment
- On Goodhart's law, with an application to value alignment
- Logical Induction
- Goodhart's Law in Reinforcement Learning
- Penalizing side effects using stepwise relative reachability
- Correlated Proxies: A New Definition and Improved Mitigation for Reward Hacking
- Categorizing Variants of Goodhart's Law
- Consequentialist Objectives and Catastrophe
- Preferences Implicit in the State of the World
- Goal Misgeneralization: Why Correct Specifications Aren't Enough For Correct Goals
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