Gacha Decoding：通过指令遵循激发多样化生成

中文摘要 AI 辅助

我们引入了Gacha Decoding，一种在推理时通过指令遵循来激发多样化语言模型生成的方法，其效果随模型能力提升而扩展。在开放领域（如野外聊天、创意写作、图像生成规划以及蛋白质设计）中，Gacha Decoding在同等质量下显著优于现有的生成多样性方法（相比次优方法，Vendi分数最高提升2.4倍），以超过一个数量级更少的样本（11.0倍）达到相同数量的高质量模式，并发现了其他方法未能触及的新模式。我们的关键洞察在于将多样性视为一个指令遵循问题：不依赖语言模型的令牌熵，而是将其指令遵循能力与外部随机数生成器（RNG）工具的随机性相结合，以可扩展的方式识别并实现响应空间中的不同模式。这种“掷骰子规划”的方法使Gacha能够逆转长期观察到的多样性与模型能力之间的张力。随着底层语言模型成为更好的指令遵循者，在Gacha Decoding下的多样性持续提升——即使其令牌熵和先前方法下的多样性下降。综合来看，我们的结果强调指令遵循，而非仅令牌熵，可以驱动生成多样性。

英文摘要

We introduce Gacha Decoding, an inference-time method for eliciting diverse language model generations that scales with model capability. Across open-ended domains (in-the-wild chat, creative writing, planning for image generation, and protein design), Gacha Decoding significantly outperforms existing generation diversity approaches at equal quality (up to 2.4x Vendi over the next-best prior approach), reaching the same number of high-quality modes with over an order of magnitude fewer samples (11.0x) and discovering novel modes that no other approach surfaces. Our key insight is to treat diversity as an instruction-following problem: rather than relying on the LM's token entropy, we combine its instruction-following capability with randomness from an external RNG tool to scalably identify and realize distinct modes of the response space. This approach of "planning with dice" enables Gacha to invert the long-observed tension between diversity and model capability. As the underlying LM becomes a better instruction follower, diversity under Gacha Decoding consistently improves--even as its token entropy and diversity under prior approaches decline. Together, our results highlight that instruction following, rather than token entropy alone, can drive generation diversity.

发表机构

• University of Washington（华盛顿大学）
• Meta Superintelligence Labs（Meta超级智能实验室）

机构由 AI 辅助整理，请以论文原文为准。