Cheap Verifiers, Large Blind Spots: Measuring the Reliability Cost of Cost-Saving Cascades
cs.AI, cs.CR, cs.LG
Submitted: 2026-09-01
Updated: 2026-09-04
Code: https://github.com/AltSlate-Labs/cascade-blindspot
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance
- Language Model Cascades
- From Deferral to Learning: Online In-Context Knowledge Distillation for LLM Cascades
- Inference-Time Distillation: Cost-Efficient Agents Without Fine-Tuning or Manual Prompt Engineering
- Accelerating Large Language Model Decoding with Speculative Sampling
- Categorizing Variants of Goodhart's Law
- RouteLLM: Learning to Route LLMs with Preference Data
- Reinforced Self-Training (ReST) for Language Modeling
- Weak-to-Strong Generalization: Eliciting Strong Capabilities With Weak Supervision
- The Limits of Inference Scaling Through Resampling
- Training Verifiers to Solve Math Word Problems
Related papers
- 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
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems
- VSAL: A Vision Solver with Adaptive Layouts for Graph Property Detection