Pith. sign in

REVIEW 2 cited by

Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference?

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.05079 v2 pith:4VKVPDTR submitted 2023-10-08 cs.LG

classification cs.LG
keywords quantisationllmsdensitytimesarithmeticmemoryscalingblock
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
abstract

The inference of Large language models (LLMs) requires immense computation and memory resources. To curtail these costs, quantisation has merged as a promising solution, but existing LLM quantisation mainly focuses on 8-bit. In this work, we explore the statistical and learning properties of the LLM layer and attribute the bottleneck of LLM quantisation to numerical scaling offsets. To address this, we adapt block quantisations for LLMs, a family of methods that share scaling factors across packed numbers. Block quantisations efficiently reduce the numerical scaling offsets solely from an arithmetic perspective, without additional treatments in the computational path. Our nearly-lossless quantised 6-bit LLMs achieve a $19\times$ higher arithmetic density and $5\times$ memory density than the float32 baseline, surpassing the prior art 8-bit quantisation by $2.5\times$ in arithmetic density and $1.2\times$ in memory density, without requiring any data calibration or re-training. We also share our insights into sub-8-bit LLM quantisation, including the mismatch between activation and weight distributions, optimal fine-tuning strategies, and a lower quantisation granularity inherent in the statistical properties of LLMs. The latter two tricks enable nearly-lossless 4-bit LLMs on downstream tasks. Our code is open-sourced.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Knowledge Distillation of Domain-adapted LLMs for Question-Answering in Telecom

    cs.CL 2025-04 conditional novelty 6.0 of 10

    Fine-tuning the teacher before knowledge distillation improves the distilled model most reliably when teacher and student share a vocabulary, and fine-tuning both gives the best scores overall.

  2. BBAL: A Bidirectional Block Floating Point-Based Quantisation Accelerator for Large Language Models

    cs.AR 2025-04 conditional novelty 5.0 of 10

    A flag-bit block floating point format reduces quantization error versus BFP and enables a hardware accelerator for LLM linear and nonlinear layers.

Pith tools