❌

Reading view

RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M

A reproducible test of BM25, embeddings, and parent documents shows why author-written evals overrate search, and why answer position in context matters.

The post RAG Retrieval Lost 40 Points When Distracting Passages Grew to 2.68M appeared first on Towards Data Science.

  •  

β€œAre You Sure?” Flipped Qwen2.5-1.5B’s Right Answer 2 Times in 3

Three small open models, 150 MMLU questions each, five pushback styles: β€œThat’s wrong” did more damage, and naming a wrong option was close to fatal.

The post β€œAre You Sure?” Flipped Qwen2.5-1.5B’s Right Answer 2 Times in 3 appeared first on Towards Data Science.

  •  

Nebius and Positronic launch physical AI competition to test robot manipulation models

Positronic Robotics and Nebius have launched an international physical AI competition aimed at improving robotic manipulation and establishing more consistent methods for evaluating AI models on real-world tasks. Announced on October 26, the Nebius Γ— Positronic Physical AI Competition runs through December 2026, inviting developers and research teams to submit robotic AI models for independent […]
  •  

I Told an AI Agent That Its Wrong Answers Were Right. One Model Believed Me.

A research paper claims reward labels barely change how agents learn. I tested that on three models, and one of them did exactly what the wrong labels told it to.

The post I Told an AI Agent That Its Wrong Answers Were Right. One Model Believed Me. appeared first on Towards Data Science.

  •  

An AI Agent Can Spot a Mistake and Still Not Fix It

I replayed correct model decisions through four deliberately broken browser actions, and a simple page check missed half of the wrong clicks, which shows that noticing a fault and recovering from it are different skills.

The post An AI Agent Can Spot a Mistake and Still Not Fix It appeared first on Towards Data Science.

  •  

Does Agentic RAG Actually Retrieve Better Evidence? I Tested It on 298 Questions

I let an AI agent search again whenever it chose. It found some papers plain search had missed, lost some plain search had found, and gained too little to trust.

The post Does Agentic RAG Actually Retrieve Better Evidence? I Tested It on 298 Questions appeared first on Towards Data Science.

  •  
  •  

Coding Agents Can Solve the Task β€” But How Much Work Did They Waste?

Real repositories, seeded agent decisions: what 50,400 runs show about wasted turns, patch churn, and which harness rules help

The post Coding Agents Can Solve the Task β€” But How Much Work Did They Waste? appeared first on Towards Data Science.

  •  
❌