pith. sign in

• (Errica et al., 2024) show that, even when average accuracy is high, minor changes 16 Table 13: Accuracy achieved by prompts from RLPrompt and PRL on classification tasks

show that, even for instruction-tuned chat models, the gap between the best, worst semantically equivalent prompts remains large, that small rephrasings can significantly alter model behaviour · 2024

1 Pith paper cite this work. Polarity classification is still indexing.

1 Pith paper citing it
browse 1 citing papers

fields

cs.AI 1

years

2025 1

verdicts

UNVERDICTED 1

representative citing papers

PRL: Prompts from Reinforcement Learning

cs.AI · 2025-05-20 · unverdicted · novelty 5.0

PRL is a reinforcement learning method that generates novel prompts and achieves state-of-the-art results on text classification, simplification, and summarization benchmarks, outperforming APE and EvoPrompt.

citing papers explorer

Showing 1 of 1 citing paper.

  • PRL: Prompts from Reinforcement Learning cs.AI · 2025-05-20 · unverdicted · none · ref 16

    PRL is a reinforcement learning method that generates novel prompts and achieves state-of-the-art results on text classification, simplification, and summarization benchmarks, outperforming APE and EvoPrompt.