# Nacrith: Neural Lossless Compression via Ensemble Context Modeling and High-Precision CDF Coding

arXiv: https://arxiv.org/abs/2602.19626 · HTML: https://mlsanity.com/paper/2602.19626/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 5 of its 26 references that are in the map.
- First posted on arXiv 23 Feb 2026.

## Builds on (most cited first, 5 of 5)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Language Modeling Is Compression](https://mlsanity.com/paper/2309.10668/) (2023)
- [LLMZip: Lossless Text Compression using Large Language Models](https://mlsanity.com/paper/2306.04050/) (2023)
- [FineZip : Pushing the Limits of Large Language Models for Practical Lossless Text Compression](https://mlsanity.com/paper/2409.17141/) (2024)
- [Llamazip: Leveraging LLaMA for Lossless Text Compression and Training Dataset Detection](https://mlsanity.com/paper/2511.17589/) (2025)

## Built on by (most cited first, 4 of 4)

- [Micro-Diffusion Compression - Binary Tree Tweedie Denoising for Online Probability Estimation](https://mlsanity.com/paper/2603.08771/) (2026)
- [StateSMix: Online Lossless Compression via Mamba State Space Models and Sparse N-gram Context Mixing](https://mlsanity.com/paper/2605.02904/) (2026)
- [Cadence: Error-Bounded Lossy Compression of Demand Time Series with a Time-Series Foundation Model](https://mlsanity.com/paper/2609.06008/) (2026)
- [Lossy Compressive Text Autoencoders](https://mlsanity.com/paper/2610.10738/) (2026)

## Abstract (arXiv)

We present Nacrith, a lossless compression system that combines a 135M-parameter transformer language model (SmolLM2-135M) with an ensemble of lightweight online predictors and a 32-bit arithmetic coder, achieving the best compression results among the systems evaluated in this study on natural language text. Beyond the base LLM-plus-arithmetic-coding paradigm, Nacrith introduces several contributions: (1) a CDF precision upgrade from 2^16 to 2^24 that eliminates ~75% of quantization overhead caused by minimum-probability floors in large vocabularies; (2) a token-level N-gram model for fast local predictions; (3) an adaptive log-space bias head correcting per-document LLM errors via online gradient descent; (4) confidence-based LLM skip for accelerating highly predictable tokens; (5) a hybrid binary format (NC06) extending neural compression to arbitrary binary files--to our knowledge a first among LLM-based compressors; (6) a llama cpp inference backend achieving ~7x faster single-token decode than PyTorch; (7) parallel multi-GPU compression across up to 8 workers; and (8) native KV cache sliding window reducing per-slide cost by ~37x. The system requires only ~500 MB of GGUF weights and ~1.2 GB VRAM per worker, running on consumer GPUs. On alice29 (Canterbury Corpus, 152 KB), Nacrith achieves 0.918 bits per byte (bpb)--outperforming gzip by 3.1x, bzip2 by 2.5x, CMIX v21 by 44%, and ts_zip by 20%, while compressing below the 0th-, 1st-, and 2nd-order byte-level Shannon entropy bounds. On enwik8 (100 MB), Nacrith achieves 0.9389 bpb (11.74%), surpassing ts_zip (~1.11 bpb) by 15% and FineZip (1.024 bpb) by 8% despite using a 60x smaller model with no fine-tuning. An out-of-distribution (OOD) evaluation on a document published after the model's training cutoff confirms these gains are not memorization artifacts, achieving 0.723 bpb on unseen text.