arXiv Machine Learning July 7, 2026 By Teng-Ruei Chen

How Much of the Routing Gap Is Real? Decomposing the Router-to-Oracle Gap into Reproducible Specialist Advantage and Single-Draw Label Noise

Read the original on arXiv Machine Learning →

arXiv:2607. 03436v1 Announce Type: new Abstract: Routing among large language models (LLMs) promises better quality at lower cost, motivated by the reported gap between learned routers and a per-instance oracle.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

arXiv Machine Learning

Aug 11

Opportunity Is Not Realizability: Selection-Valid Diagnostics for Multi-LLM Routing

arXiv:2608. 08265v1 Announce Type: new Abstract: Oracle routing measures how much a pool of language models could gain from per-query selection, but the diagnostic has two flaws: testing against a best fixed model selected on the same examples invalidates paired inference, and a full-information oracle sees outcomes no deployable router observes.

By Ibne Farabi Shihab, Abu Sa-Adat Mohamed Moon-Im Al Ahsan, Md Najmus Swaqeeb

llms benchmarks

arXiv Machine Learning

Jul 10

Resample or Reroute? Budget-Aware Test-Time Model Selection for Large Language Models

arXiv:2607. 08665v1 Announce Type: new Abstract: Routing among large language models (LLMs) trades response quality against serving cost, motivated by the reported gap between deployed routers and a per-instance oracle.

By Teng-Ruei Chen

llms benchmarks

arXiv Machine Learning

Aug 12

Detecting an Effect Is Not Learning to Act on It: A Reward-SNR Floor for LLM Acquisition Agents

arXiv:2608. 10441v1 Announce Type: new Abstract: Many pipelines can pay a per-example cost to acquire an auxiliary, model-derived observation -- an LLM's structured reasoning, a slow oracle, an expensive measurement -- and then must decide when the acquired signal is worth using.

By Ying Yuan

llms rag agents

arXiv Machine Learning

Jun 2

Auditing Near-Optimal Policies Can Be Exponentially Hard: Conditional Query Lower Bounds via Occupancy Rashomon Capacity

arXiv:2606. 00414v1 Announce Type: new Abstract: When many reinforcement-learning policies achieve near-optimal return, a post-hoc auditor may have to distinguish among many behaviorally distinct but return-equivalent policies.

By Ibne Farabi Shihab, Sanjeda Akter, Anuj Sharma

reinforcement-learning benchmarks

arXiv Machine Learning

Jun 25

Reclaim Evaluation: A Lossy Memory Is Worse Than an Empty One

arXiv:2606. 25449v1 Announce Type: cross Abstract: A language model's memory can be worse than having no memory at all.

By Alex Kwon

llms benchmarks

arXiv AI

Jun 16

Entropy-Gated Latent Recursion

arXiv:2606. 16620v1 Announce Type: cross Abstract: Inference-time scaling has become the dominant lever for improving language-model reasoning, but existing methods derive rollout diversity from a single source: stochastic token-level sampling.

By Soham Bhattacharjee, Dushyant Singh Chauhan, Salem Lahlou, Martin Takac, Nils Lukas

reinforcement-learning benchmarks

Related stories

Opportunity Is Not Realizability: Selection-Valid Diagnostics for Multi-LLM Routing

Resample or Reroute? Budget-Aware Test-Time Model Selection for Large Language Models

Detecting an Effect Is Not Learning to Act on It: A Reward-SNR Floor for LLM Acquisition Agents

Auditing Near-Optimal Policies Can Be Exponentially Hard: Conditional Query Lower Bounds via Occupancy Rashomon Capacity

Reclaim Evaluation: A Lossy Memory Is Worse Than an Empty One

Entropy-Gated Latent Recursion