Stanford researchers just developed Quiet-STaR, a new training method that enables AI models to ‘think’ before responding.

When applied to the Mistral 7B language model, Quiet-STaR nearly doubled its scores on math and common sense reasoning benchmarks. https://t.co/jv02P5yStX

— Rowan Cheung (@rowancheung) 22 mars 2024

Stanford researchers just developed Quiet-STaR, a new training method that enables AI models to ‘think’ before responding. When applied to the Mistral 7B language model, Quiet-STaR nearly doubled its scores on math and common sense reasoning benchmarks.

→ View original post on X — @rowancheung,