StateBench tests how language models track changes in long documents
A new benchmark called StateBench has been presented on Hugging Face. It is designed to test how well language models track state changes over long documents.
More news
- AI agents search for alien intelligence at SETI's telescope2026-11-04
- Google Antigravity helps speed up drug discovery2026-11-04
- New Apple Watch beta lets Siri listen in and summarize conversations2026-11-04
- Anthropic bans cloning existing games with Claude Code2026-11-04
- OpenAI launches program to support power grids2026-11-03
- Anthropic grows revenue twelvefold but posts $23 billion loss2026-11-03
- 42AI emerges from stealth with $62.5M2026-11-03
- Agentic analytics: the problem of AI agents' insight quality2026-11-03