TinyCeNN-LM proposes a quality‑gated post‑training conversion framework that replaces attention in pretrained language models with CeNN‑inspired cellular‑recurrent layers. The method includes bounded local processing, compact recurrent memory, routing, fusion, and an accept‑or‑rollback validation step, with three specific implementations studied. Experiments on SmolLM2‑135M and Qwen3.5‑0.8B show that the conversion can accept certain layers while rejecting others based on representation fidelity and NLL thresholds, achieving minimal perplexity changes and modest downstream accuracy retention.
arXiv:2608.30320v1 Announce Type: new Abstract: We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and a...
arXiv:2607. 07953v1 Announce Type: cross Abstract: Self-attention lets each token retrieve information from the full context, but its quadratic cost in sequence length limits training and inference at long context.
The paper introduces dQwen3.5, a family of diffusion language models derived from the hybrid-attention architecture of Qwen3.5 at 0.8B, 2B, 4B, and 9B parameters. It demonstrates that adapting a hybrid backbone—combining attention and RNN layers—can be more efficient than full-attention models, reaching a target training loss in roughly half the tokens. Across scales, dQwen3.5 exhibits full-attention-like behavior in any-order decoding and strong performance with parallel decoding.
The paper reports a large‑scale post‑training ternarisation of the Qwen3 language model, extending a conversion pipeline from the 4B to the 8B variant. Using KOTMS rotation, E2M‑ATQ adaptive ternarisation, and GPTQ‑style error compensation, the authors achieve a 1.361× perplexity ratio across three corpora and retain 78.5% of the FP16 accuracy on zero‑shot tasks, with the 8B model outperforming the 4B by 8.9 percentage points. The study also demonstrates lossless lattice‑aware packing, producing an 8.24 GiB checkpoint that preserves perplexity, and shows that direct packed execution can reach 15.52 tokens/s in 7.35 GiB, though packed GEMV remains slower than FP16 cuBLAS.
The paper presents a 4‑bit quantization recipe, Minima: NVFP4 W4A4, that fully quantizes all linear layers—including the Gated DeltaNet (GDN) recurrent blocks—of the 27‑billion‑parameter Qwen3.8 LLM. Across a suite of benchmarks (perplexity, MMLU‑Pro, GSM8K, AIME'25, GPQA‑Diamond, LiveCodeBench, and RULER retrieval), the quantized model matches BF16 performance within seed noise while being 17.5 GiB in size and 14–19 % faster at prefill. The authors attribute this success to four mechanisms: block‑scaling of residuals, robust gate projections, the delta‑rule recurrence’s noise‑plateau behavior, and the per‑token quantization cost’s dilution over long contexts.
The paper introduces dQwen3.5, a family of diffusion language models derived from the hybrid-attention architecture of Qwen3.5 at 0.8B, 2B, 4B, and 9B parameters. It demonstrates that adapting a hybrid AR backbone—combining attention and RNN layers—can be more efficient than full-attention models, reaching a target training loss in roughly half the tokens. Across scales, dQwen3.5 exhibits full-attention-like behavior in any‑order decoding and strong performance under parallel decoding.
arXiv:2608. 20210v1 Announce Type: cross Abstract: Small language models are usually built like large ones and then squeezed onto a CPU afterwards.
The paper reports a post‑training ternarization of the 4‑billion‑parameter Qwen model, achieving an effective 1.641‑bit representation for 81.62 % of its weights while keeping activations at 16‑bit precision. Accuracy drops from 64.5 % to 54.7 % across ten capability tests, with uneven degradation (e.g., BoolQ 84.6 % of teacher performance, ARC‑Challenge 43.8 %). After packing the ternary planes, the model size shrinks from 8.29 GiB to 3.96 GiB with negligible change in perplexity, though inference speed is not improved.
The paper investigates how block‑diffusion language models can use a constant‑size cache to enable efficient parallel decoding. By employing sequence mixers that summarize completed blocks into a reusable state and a block‑causal training objective, the authors pretrain three 3B block‑diffusion denoisers (attention, Mamba, and hybrid) on 300 B tokens. The resulting state‑space cache remains O(1) in memory and latency regardless of context length, yielding significant speed‑up and memory savings compared to traditional attention‑based caches, especially at very long sequences.
Small language models are usually built like large ones and then squeezed onto a CPU afterwards. We did the opposite: we fixed the target first, one user, one token at a time, 4-bit weights, ordinary CPU, and chose the architecture to suit it.
The paper introduces LatentPort, a method that allows a language model to transfer its live memory to another model without requiring the receiver to reread the context. Experiments on a Qwen3.5 4B-to-9B sibling pair show that adding a Gated DeltaNet (GDN) persistent-state package reduces negative log‑likelihood by 0.747 nats/token and improves performance across 64 PG19 documents. The study also demonstrates that direct recurrent and convolution reuse outperforms learned GDN maps, and a 434,176‑parameter correction further narrows the performance gap to the native 9B model.
arXiv:2506. 21833v2 Announce Type: replace Abstract: Forward-mode automatic differentiation (FmAD) and zero-order (ZO) optimization are increasingly proposed as memory-efficient, backpropagation-free alternatives for large language model (LLM) fine-tuning, yet their benefits are typically evaluated only against standard backpropagation (BP), omitting memory-efficient variants such as activation checkpointing.