Tag
Proposes a tree-of-thoughts inspired extractive-abstractive approach for legal case judgement summarization using LLMs, with experiments on DeepSeek and LLama showing improved summaries over extractive or abstractive methods alone.
The author introduces a novel approach for AI agents called ADHD, using parallel divergent ideation to enable non-linear thinking inspired by ADHD minds, though with higher cost and latency, and releases it as open source.
A researcher proposes a method called 'ADHD - Parallel Divergent Ideation' that uses tree-of-thoughts to give AI agents divergent thinking, inspired by ADHD, but with significant trade-offs in cost and speed. The project is open-source.