Quantifying Behavioral Tails in Black-Box Language Models
cs.LG, cs.CL, stat.ML
Submitted: 2026-09-27
Updated: 2026-09-29
Terminology
Sources
- Phi-4-reasoning Technical Report
- Estimating Tail Risks in Language Model Output Distributions
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
- Rare Event Analysis of Large Language Models
- Mistral 7B
- Forecasting Rare Language Model Behaviors
- OverThink: Slowdown Attacks on Reasoning LLMs
- Shared Global and Local Geometry of Language Model Embeddings
- POT: Inducing Overthinking in LLMs via Black-Box Iterative Optimization
- ThinkTrap: Denial-of-Service Attacks against Black-box LLM Services via Infinite Thinking
- 2 OLMo 2 Furious
- gpt-oss-120b & gpt-oss-20b Model Card
- Estimating Rare Events in Language Models with Proper Evaluation
- SCARCE: Scalable Cascade Analysis for Rare-event Characterisation via Embeddings
- Predicting LLM Safety Before Release by Simulating Deployment
- Qwen3 Technical Report
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks