Blog

Beyond Next-Token Prediction: How Post-Training Teaches LLMs to Reason

Agent Evaluation: Why Simulated Environments are the New Frontier for Data

Toxicity detection: Why we still need human-labeled data

Evaluating Model Reasoning with Rubrics: Building a Domain-Specific Evaluation Dataset

Human-powered evaluation: Actionable feedback for next‑gen video diffusion models

Standardizing AI safety with MLCommons

Toloka Fuels Next Stage of Growth with Investment Led by Bezos Expeditions

AI agents under attack: A case study on advanced agent red-teaming

Introducing JEEM: A new benchmark for evaluating low-resource Arabic dialects

The personality paradox: Teaching AI agents to act like real people

Fixing SWE-bench: A Smarter Way to Evaluate Coding AI

LLM Evaluation In Action: Should You Trust Automated Metrics or Human Judgment?

Toloka’s Commitment to Responsible AI: How We Prioritize Ethics, Safety, and Excellence

Introducing Toloka’s Bug Bounty Program: Strengthening Security with Ethical Hacking

R1 is not on par with o1, and the difference is qualitative, not quantitative