Penalty Decoding: Well Suppress the Self-Reinforcement Effect in Open-Ended Text Generation
Submission Type: Regular Short Paper
Submission Track: Natural Language Generation
Submission Track 2: Language Modeling and Analysis of Language Models
Keywords: decoding algorithm, open-ended text generation, self-reinforcement, repetition penalty
TL;DR: This paper analyzes the phenomenon of self-reinforcement in open text generation task and improves the existing problem of repetition penalty.
Abstract: The decoding algorithm is critical for open-ended text generation, transforming latent representations into coherent and meaningful outputs. This paper investigates the self-reinforcement effect in text generation and the effectiveness of a repetition penalty to mitigate it. However, determining the optimal repetition penalty value is challenging. To tackle this, we propose a forgetting mechanism that disregards distant tokens, reducing the burden of penalty selection. In addition, we introduce a length penalty to address overly short sentences caused by excessive penalties. Our penalty decoding approach incorporating three strategies helps resolve issues with sampling methods deviating from factual information. Experimental results demonstrate the efficacy of our approach in generating high-quality sentences resembling human output.
Submission Number: 2527
Loading