The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context
  Learning

Bill Yuchen Lin; Abhilasha Ravichander; Ximing Lu; Nouha Dziri; Melanie Sclar; Khyathi Chandu; Chandra Bhagavatula; Yejin Choi

Paper · arXiv 2312.01552

The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning

Bill Yuchen Lin, Abhilasha Ravichander, Ximing Lu, Nouha Dziri, Melanie Sclar, Khyathi Chandu, Chandra Bhagavatula, Yejin Choi

31 upvotesDecember 4, 2023arXiv 预印本

AI 摘要

A study explores the effectiveness of alignment tuning in large language models, suggesting that superficial methods like in-context learning can achieve strong performance without traditional supervised fine-tuning or reinforcement learning.

instruction learningsupervised fine-tuningreinforcement learning from human feedbacktoken distribution shiftstylistic tokensSuperficial Alignment HypothesisURIALin-context learningJUST-EVAL-INSTRUCTtuning-free alignment

Abstract

The alignment tuning process of large language models (LLMs) typically involves instruction learning through supervised fine-tuning (SFT) and preference tuning via reinforcement learning from human feedback (RLHF). A recent study, LIMA (Zhou et al. 2023), shows that using merely 1K examples for SFT can achieve significant alignment performance as well, suggesting that the effect of alignment tuning might be "superficial." This raises questions about how exactly the alignment tuning transforms a base LLM. We analyze the effect of alignment tuning by examining the token distribution shift between base LLMs and their aligned counterpart. Our findings reveal that base LLMs and their alignment-tuned versions perform nearly identically in decoding on the majority of token positions. Most distribution shifts occur with stylistic tokens. These direct evidence strongly supports the Superficial Alignment Hypothesis suggested by LIMA. Based on these findings, we rethink the alignment of LLMs by posing the research question: how effectively can we align base LLMs without SFT or RLHF? To address this, we introduce a simple, tuning-free alignment method, URIAL. URIAL achieves effective alignment purely through in-context learning (ICL) with base LLMs, requiring as few as three constant stylistic examples and a system prompt. We conduct a fine-grained and interpretable evaluation on a diverse set of examples, named JUST-EVAL-INSTRUCT. Results demonstrate that base LLMs with URIAL can match or even surpass the performance of LLMs aligned with SFT or SFT+RLHF. We show that the gap between tuning-free and tuning-based alignment methods can be significantly reduced through strategic prompting and ICL. Our findings on the superficial nature of alignment tuning and results with URIAL suggest that deeper analysis and theoretical understanding of alignment is crucial to future LLM research.