Metraipolis › archive › dc_a1aaae91a35d

Large Language Models with At Most One Spike per Neuron — arXiv:2609.05151

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.05151
License (as archived)
CC0-1.0
Archived
2026-10-16T22:36:04.607Z
SHA-256 of the archived text
479a734a2134f09191254f62d179cd4f365a744b5d25c45b4b632c2ab0324701
Stable URL of this page
https://metraipolis.com/documento/dc_a1aaae91a35d

How to cite

Large Language Models with At Most One Spike per Neuron — arXiv:2609.05151 — arXiv (en), https://arxiv.org/abs/2609.05151 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a1aaae91a35d (Metraipolis archive dc_a1aaae91a35d, sha256 479a734a2134f09191254f62d179cd4f365a744b5d25c45b4b632c2ab0324701)
BibTeX
@misc{dca1aaae91a35d,
  title = {Large Language Models with At Most One Spike per Neuron — arXiv:2609.05151},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.05151},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a1aaae91a35d; sha256 479a734a2134f09191254f62d179cd4f365a744b5d25c45b4b632c2ab0324701},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.05151. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Leveraging their inherent sparse event-driven computation, spiking neural networks (SNNs) offer a promising path toward energy-efficient large language models (LLMs). Time-to-first-spike (TTFS) coding generates at most one spike per neuron within a time window, yielding extremely low firing rates. However, conventional TTFS SNNs are restricted to specific structures, making it challenging to encode certain blocks in LLM -- such as layer normalization and matrix multiplication --using TTFS. To overcome this limitation, we introduce a reference-based strategy specifically to encode the four core LLM components: embedding layers, layer normalization, attention-related operations and dropout. We construct a fully TTFS-based SNN architecture and train it end-to-end. Experiments on modern LLMs like BERT and GPT-2 demonstrate that our approach achieves performance comparable to ANN counterparts on natural language understanding and common-sense reasoning, while a clear gap remains on language modeling perplexity. To the best of our knowledge, this is the first work to scale a spiking LLM to 1.5 billion parameters using TTFS coding. We also report an estimate of spike-related energy; this is a spike-count proxy under an established cost model rather than a measurement on neuromorphic hardware.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-07 in cs.CL).
Categorie arXiv: Neural and Evolutionary Computing (cs.NE) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.05151 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T22:21:16.326Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).