iFairy: the First 2-bit Complex LLM with All Parameters in $\{\pm1, \pm i\}$

TL;DR

This paper introduces Fairy±i, a novel 2-bit complex-valued LLM quantization method that surpasses existing accuracy ceilings by leveraging complex domain advantages and symmetric representations, enabling efficient, multiplication-free inference.

Contribution

It proposes the first 2-bit complex-valued LLM quantization framework that exceeds the accuracy of full-precision models by utilizing complex domain representations and symmetric weight mappings.

Findings

Outperforms existing 2-bit quantization methods in PPL and downstream tasks.

Enables multiplication-free inference with addition and swaps.

Maintains strict storage and compute efficiency.

Abstract

Quantization-Aware Training (QAT) integrates quantization into the training loop, enabling LLMs to learn robust low-bit representations, and is widely recognized as one of the most promising research directions. All current QAT research focuses on minimizing quantization error on full-precision models, where the full-precision accuracy acts as an upper bound (accuracy ceiling). No existing method has even attempted to surpass this ceiling. To break this ceiling, we propose a new paradigm: raising the ceiling (full-precision model), and then still quantizing it efficiently into 2 bits. We propose Fairy$\pm i$, the first 2-bit quantization framework for complex-valued LLMs. Specifically, our method leverages the representational advantages of the complex domain to boost full-precision accuracy. We map weights to the fourth roots of unity $\{\pm1, \pm i\}$, forming a perfectly symmetric and information-theoretically optimal 2-bit representation. Importantly, each quantized weight has either a zero real or imaginary part, enabling multiplication-free inference using only additions and element swaps. Experimental results show that Fairy$\pm i$ outperforms the ceiling of existing 2-bit quantization approaches in terms of both PPL and downstream tasks, while maintaining strict storage and compute efficiency. This work opens a new direction for building highly accurate and practical LLMs under extremely low-bit constraints.