learn.ifs.name

ReAct: Synergizing Reasoning and Acting in Language Models

Yao et al., 2023, ReAct (The Eleventh International Conference on Learning Representations (ICLR 2023))

Subscriber deck — free during early access

2 cards in “3.2 Methods” · whole deck

3.2 Methods

  1. What baselines did ReAct’s authors derive by ablating its trajectories?
    Standard prompting (no thoughts, actions or observations), chain-of-thought (thoughts only), CoT with self-consistency (21 samples, majority vote) and Act (actions and observations only, without thoughts).
    • Yao et al., 2023, ReAct (The Eleventh International Conference on Learning Representations (ICLR 2023)), §3.2
  2. How did Yao et al. combine ReAct with CoT-SC?
    Two back-off heuristics: ReAct → CoT-SC falls back to self-consistency when ReAct gives no answer within a step limit (7 for HotpotQA, 5 for FEVER); CoT-SC → ReAct switches to ReAct when the majority answer gets fewer than half of the samples, i.e. internal knowledge seems unsure.
    • Yao et al., 2023, ReAct (The Eleventh International Conference on Learning Representations (ICLR 2023)), §3.2

Source

  • Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan and Yuan Cao. 2023. ReAct: Synergizing Reasoning and Acting in Language Models. In The Eleventh International Conference on Learning Representations (ICLR 2023). https://arxiv.org/abs/2210.03629