AI Paper Atlas

Research map / Architectures and efficiency / Quantization

BBQ: Boosting Quantization Entropy with Bell Box Quantization

Ningfeng Yang, Tor M. Aamodt

ICLR 2026 · Architectures and efficiency › Quantization · approach: Mixed-precision quantization

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Hadamard Transform (ablation)BBQ · 2-bit LLaMA-95M, 3B tokensperplexity: 35.79 -> 31.34Improved
add RMS normalization (ablation)BBQ · 2-bit LLaMA-95M, 3B tokensperplexity: 35.93 -> 31.34Improved
substitute clip with probability integral transform without gamma init (ablation)QuEST · 2-bit LLaMA-95M, 3B tokensperplexity: 35.58 -> 138.3Hurt
add PIT together with gamma initialization instead of clip (ablation)QuEST · 2-bit LLaMA-95M, 3B tokensperplexity: 35.58 -> 31.46Improved
make gamma learnable (ablation)BBQ · 2-bit LLaMA-95M, 3B tokensperplexity: 31.46 -> 31.34Improved
use BBQ instead of QuEST at 1-bitQuEST · C4 LLaMA-95M W1A1perplexity: 67.78 -> 49.22Improved
use BBQ instead of QuEST at 2-bitQuEST · C4 LLaMA-200M W2A2perplexity: 25.46 -> 23.08Improved
use BBQ instead of QuEST at 4-bitQuEST · C4 LLaMA-125M W4A4perplexity: 22.98 -> 22.15Improved
use BBQ instead of LSQ at 2-bitLSQ · C4 LLaMA-95M W2A2perplexity: 36.58 -> 31.34Improved
use BBQ instead of QuEST at 2-bitQuEST · C4 GPT 95M W2A2perplexity: 36.7 -> 34.36Improved
use BBQ instead of N2UQ at 2-bitN2UQ · C4 GPT 95M W2A2perplexity: 35.59 -> 34.36Improved
use BBQ instead of NF4 for 4-bit inferenceNF4 · LLaMA prefill latencylatency: 48% speedup over NF4Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy