Hugging Face's logo Hugging Face
  • Models
  • Datasets
  • Spaces
  • Buckets
  • Docs
  • Enterprise
  • Pricing
    • Website
      • Tasks
      • HuggingChat
      • Collections
      • Languages
      • Organizations
    • Community
      • Blog
      • Posts
      • Daily Papers
      • Hardware
      • Learn
      • Discord
      • Forum
      • GitHub
    • Solutions
      • Team & Enterprise
      • Hugging Face PRO
      • Enterprise Support
      • Inference Providers
      • Inference Endpoints
      • Storage Buckets

  • Log In
  • Sign Up

exea-labs
/
ExeaLM

English
Japanese
exealm
open-source
from-scratch
research-preview
tonn
hybrid
retrieval
multi-agent
Model card Files Files and versions
xet
Community
ExeaLM
2.54 GB
Ctrl+K
Ctrl+K
  • 1 contributor
History: 9 commits
AvnehBhatia's picture
AvnehBhatia
Overnight runs of 2026-11-12: the 18M model with gated attention and QK-norm (1.393 against 1.419), with the circuit search (1.431), the 0.1B model at step 3,000, the 8,000-step wide fine-tune, and their result files
ce2a9da verified about 9 hours ago
  • checkpoints
    Overnight runs of 2026-11-12: the 18M model with gated attention and QK-norm (1.393 against 1.419), with the circuit search (1.431), the 0.1B model at step 3,000, the 8,000-step wide fine-tune, and their result files about 9 hours ago
  • results
    Overnight runs of 2026-11-12: the 18M model with gated attention and QK-norm (1.393 against 1.419), with the circuit search (1.431), the 0.1B model at step 3,000, the 8,000-step wide fine-tune, and their result files about 9 hours ago
  • tokenizer
    Model card: 0.1B hybrid design and small-scale results; add the 32k tokenizer 2 days ago
  • .gitattributes
    1.52 kB
    initial commit 3 days ago
  • README.md
    8.52 kB
    Overnight runs of 2026-11-12: the 18M model with gated attention and QK-norm (1.393 against 1.419), with the circuit search (1.431), the 0.1B model at step 3,000, the 8,000-step wide fine-tune, and their result files about 9 hours ago