conv.

All stories
AIQuiet 11d · day 12

Researchers break 1.58-bit barrier for ternary LLM compression

New packing method reduces storage footprint of three-symbol neural networks below theoretical limit.

What to know

  • Researchers achieved sub-1.585-bit storage for ternary LLMs, breaking the previous information-theoretic ceiling via improved packing methods.
  • The work improves upon the prevailing five-trit packing standard (1.625 bits per weight) used in practical deployments.
  • The result addresses a long-standing compression bottleneck in quantized model deployment.

“Ternary Large Language Models (LLM) store every weight as one of three symbols {−1,0,+1}, so the cost of a ternary model is conventionally referenced to the information-theoretic log₂3 ≈ 1.585 bits per weight.”

Evangelos Georganas, Alexander Heinecke, Pradeep Dubey, Paper authors · arXiv preprint 2609.16338 ↗ · Sep 16, 12:00 AM

Evangelos Georganas Researcher, paper authorAlexander Heinecke Researcher, paper authorPradeep Dubey Researcher, paper author

How it unfolded 2 developments, newest first · click a bar or a number to jump articlesposts

Peak 6 pieces in 3h at Sep 16, 5 PM; 15 pieces over 12 days (3 articles · 4 posts · 8 comments) Sep 15, 11 PM — 1 piece · 1 article — Newswires 1Sep 16, 2 AM — quietSep 16, 5 AM — quietSep 16, 8 AM — quietSep 16, 11 AM — quietSep 16, 2 PM — 4 pieces · 2 articles · 2 posts — Newswires 2, Hacker News 1, Mastodon 1Sep 16, 5 PM — 6 pieces · 1 post · 5 comments — Hacker News 5, Mastodon 1Sep 16, 8 PM — 1 piece · 1 comment — Hacker News 1Sep 16, 11 PM — 1 piece · 1 comment — Hacker News 1Sep 17, 2 AM — 2 pieces · 1 post · 1 comment — Hacker News 1, Mastodon 1Sep 17, 5 AM — quietSep 17, 8 AM — quietSep 17, 11 AM — quietSep 17, 2 PM — quietSep 17, 5 PM — quietSep 17, 8 PM — quietSep 17, 11 PM — quietSep 18, 2 AM — quietSep 18, 5 AM — quietSep 18, 8 AM — quietSep 18, 11 AM — quietSep 18, 2 PM — quietSep 18, 5 PM — quietSep 18, 8 PM — quietSep 18, 11 PM — quietSep 19, 2 AM — quietSep 19, 5 AM — quietSep 19, 8 AM — quietSep 19, 11 AM — quietSep 19, 2 PM — quietSep 19, 5 PM — quietSep 19, 8 PM — quietSep 19, 11 PM — quietSep 20, 2 AM — quietSep 20, 5 AM — quietSep 20, 8 AM — quietSep 20, 11 AM — quietSep 20, 2 PM — quietSep 20, 5 PM — quietSep 20, 8 PM — quietSep 20, 11 PM — quietSep 21, 2 AM — quietSep 21, 5 AM — quietSep 21, 8 AM — quietSep 21, 11 AM — quietSep 21, 2 PM — quietSep 21, 5 PM — quietSep 21, 8 PM — quietSep 21, 11 PM — quietSep 22, 2 AM — quietSep 22, 5 AM — quietSep 22, 8 AM — quietSep 22, 11 AM — quietSep 22, 2 PM — quietSep 22, 5 PM — quietSep 22, 8 PM — quietSep 22, 11 PM — quietSep 23, 2 AM — quietSep 23, 5 AM — quietSep 23, 8 AM — quietSep 23, 11 AM — quietSep 23, 2 PM — quietSep 23, 5 PM — quietSep 23, 8 PM — quietSep 23, 11 PM — quietSep 24, 2 AM — quietSep 24, 5 AM — quietSep 24, 8 AM — quietSep 24, 11 AM — quietSep 24, 2 PM — quietSep 24, 5 PM — quietSep 24, 8 PM — quietSep 24, 11 PM — quietSep 25, 2 AM — quietSep 25, 5 AM — quietSep 25, 8 AM — quietSep 25, 11 AM — quietSep 25, 2 PM — quietSep 25, 5 PM — quietSep 25, 8 PM — quietSep 25, 11 PM — quietYesterday, 2 AM — quietYesterday, 5 AM — quietYesterday, 8 AM — quietYesterday, 11 AM — quietYesterday, 2 PM — quietYesterday, 5 PM — quietYesterday, 8 PM — quietYesterday, 11 PM — quietToday, 2 AM — quietToday, 5 AM — quietToday, 8 AM — quietToday, 11 AM — quietToday, 2 PM — quietToday, 5 PM — quiet 12
Sep 16Sep 17Sep 18Sep 19Sep 20Sep 21Sep 22Sep 23Sep 24Sep 25yesterdaynow · 7:56 PM ET
  1. 2

    Paper surfaces on Hacker News frontpage

    The arXiv preprint reaches Hacker News, gaining 141 points and sparking discussion among the community about implications for model deployment and compression techniques.

    • sounds like a perfect fit for ASIC-optimized models (where matrix ops could be supported directly in BITCOS format, potentially) & achieving record power efficiency for on-device inference.And it looks like per [0], a model needs only ~30% more weights to be at comparable quality, if quantization-aware training is done...0…

      yalokHacker News10d agoview on Hacker News ↗
    2 more of the top 3 · 8 posts in this stretch
    • > We measure the actual symbol distribution of 29 ternary LLM models and find that zeros account for up to 51.5% of all weights. Motivated by this finding, we introduce BITCOS, a simple distribution-adaptive layoutI honestly assumed that's how they already work. I have to admit that I even explained it like that to a friend. Why on earth wouldn't…

      c7bHacker News10d agoview on Hacker News ↗
    • So they get down from 1.58 to 1.48 bits per weight by exploiting the fact that actual weights in practice are 0 51% of the time. Neat.If ternary llms work out and are baked into hardware as custom silicon I bet they'll be shockingly efficient.

      infogulchHacker News11d agoview on Hacker News ↗
    all of them →
  2. 1

    Paper on ternary LLM compression published on arXiv

    Evangelos Georganas, Alexander Heinecke, and Pradeep Dubey post a preprint describing a method to achieve sub-1.585-bit storage for ternary models, improving over the conventional 1.625 bits per weight using five-trit packing.

Also covered reported alongside — the timeline has no entry for these yet

  1. first by HN Best, 11d ago · also HN Frontpage, arXiv cs.AI

What people are saying 5 voices from 1 site · best of 8 · verbatim