Google's SRL framework provides a step-by-step "curriculum" that makes LLMs more reliable for complex reasoning tasks.
A tool called AI-Newton can derive scientific laws from raw data, but is some way from developing human-like reasoning.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback