新词学习作为模型引导的参数高效替代方案
Neologism Learning as a Parameter-Efficient Alternative to Fine-Tuning for Model Steering
AI总结：
本文提出通过新词学习实现模型引导，相比微调更高效，且在相同训练条件下表现更优，同时发现模型会自动生成新词。
摘要：
在语言建模中，新词是新训练的token，用于表示模型词汇中未包含的概念。新词可用于鼓励模型特定行为，例如通过添加提示'给我一个新词回答'。行为引导也可以通过微调实现，但需要更多计算资源和较少灵活性：学习新词仅训练d个参数，并允许用户仍能访问模型的默认行为。我们比较了新词学习与低秩适应（LoRA）微调的性能，发现在匹配的训练设置（相同数据和超参数）下，新词优于微调模型。我们还研究了新词的自我生成，观察到当被问及新词时，模型偶尔会自己创造新的词汇。
英文摘要：
In language modeling, neologisms are new tokens trained to represent a concept not already included in a given model's vocabulary. Neologisms can be used to encourage specific behavior in models, for example by appending prompts with "Give me a neologism answer." Behavioral steering can also be achieved through fine-tuning, albeit with more compute and less flexibility: learning a neologism only trains d parameters and allows the user to still access the model's default behavior. We compare the performance of neologism learning against low-rank adaptation (LoRA) fine-tuning, finding that neologisms outperform fine-tuned models under a matched training setup (same data and hyperparameters). We also investigate self-verbalizations of neologisms, and observe that the model will occasionally make up its own new words when asked about a neologism.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
