1 paper
Zena Al-Khalili, Rafi Hakim, Dietrich Klakow +1
Parallel thinking has enjoyed great success for boosting LLM performance on reasoning tasks without the need for any re-training. However, existing methods follow a think-first-the…