Any single-output LLM ensemble is accuracy-capped at 1-beta where beta is the all-models-wrong rate, a quantity not captured by pairwise correlations and frequently underestimated by copula models.
A unified approach to routing and cascad- ing for llms.arXiv preprint arXiv:2410.10347
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 8roles
background 1polarities
background 1representative citing papers
A router that decomposes uncertainty to flexibly route queries between cheap models and oracles while providing regret bounds and supporting abstention in classification tasks with multiple annotations.
A regime theory selects the optimal controller class for LLM action decisions from a nested lattice of four classes using three data-estimable bottlenecks, with a Bernstein-tight threshold and empirical matches on multiple benchmarks.
FairTutor achieves 97.1% of premium pedagogical quality at 71.6% lower cost via equity-aware LLM routing and introduces the TutorAccessEval benchmark to measure access-tier gaps.
LatentRouter routes image-question queries to the best MLLM by predicting counterfactual performance via latent communication between learned query capsules and model capability tokens.
LLM reliability techniques are unified as communication channel operators, with a new cost-aware router achieving superior quality-cost tradeoffs on hard tasks.
AutoRelAnnotator routes queries through fine-tuned classifier cascades with isotonic calibration to deliver high-accuracy relevance labels at roughly half the compute cost while adding a small accuracy gain.
A systematic survey of LLM ensemble methods organized into a taxonomy of ensemble-before-inference, ensemble-during-inference, and ensemble-after-inference stages, with review of benchmarks, applications, and future directions.
citing papers explorer
-
When Does Combining Language Models Help? A Co-Failure Ceiling on Routing, Voting, and Mixture-of-Agents Across 67 Frontier Models
Any single-output LLM ensemble is accuracy-capped at 1-beta where beta is the all-models-wrong rate, a quantity not captured by pairwise correlations and frequently underestimated by copula models.
-
Flexible Routing via Uncertainty Decomposition
A router that decomposes uncertainty to flexibly route queries between cheap models and oracles while providing regret bounds and supporting abstention in classification tasks with multiple annotations.
-
A Regime Theory of Controller Class Selection for LLM Action Decisions
A regime theory selects the optimal controller class for LLM action decisions from a nested lattice of four classes using three data-estimable bottlenecks, with a Bernstein-tight threshold and empirical matches on multiple benchmarks.
-
FairTutor: Equity-Aware Pedagogical LLM Routing for Budget-Constrained AI Tutoring
FairTutor achieves 97.1% of premium pedagogical quality at 71.6% lower cost via equity-aware LLM routing and introduces the TutorAccessEval benchmark to measure access-tier gaps.
-
LatentRouter: Can We Choose the Right Multimodal Model Before Seeing Its Answer?
LatentRouter routes image-question queries to the best MLLM by predicting counterfactual performance via latent communication between learned query capsules and model capability tokens.
-
A Communication-Theoretic Framework for LLM Agents: Cost-Aware Adaptive Reliability
LLM reliability techniques are unified as communication channel operators, with a new cost-aware router achieving superior quality-cost tradeoffs on hard tasks.
-
AutoRelAnnotator: Calibrated Model Cascades for Cost-Efficient Relevance Evaluation in Sponsored Search
AutoRelAnnotator routes queries through fine-tuned classifier cascades with isotonic calibration to deliver high-accuracy relevance labels at roughly half the compute cost while adding a small accuracy gain.
-
Harnessing Multiple Large Language Models: A Survey on LLM Ensemble
A systematic survey of LLM ensemble methods organized into a taxonomy of ensemble-before-inference, ensemble-during-inference, and ensemble-after-inference stages, with review of benchmarks, applications, and future directions.