All pull requests

Pull requests

• Add stream lane lab example

  #924·klei22 opened Oct 20, 2026Collaborator·

  ·
  1

• Add relu2triton experiments

  #923·klei22 opened Oct 18, 2026Collaborator·

  ·
  1

• Update Bio-dashboards to Bio-dash V3.0

  #922·Meapy011 opened Oct 7, 2026·

  ·

• Add hf arch search

  #920·klei22 opened Oct 3, 2026Collaborator·

  ·
  4

• Add ngpt experiment

  #919·klei22 opened Sep 30, 2026Collaborator·

  ·
  3

• Add experiments for comparing context extension

  #918·klei22 opened Sep 29, 2026Collaborator·

  ·
  5

• Add mel pipeline

  #916·klei22 opened Sep 29, 2026Collaborator·

  ·
  6

• feat(mlp): add Hadamard channel mixer

  #913·klei22 opened Sep 27, 2026Collaborator·

  ·
  1

• Add mel mc int clean

  #910·klei22 opened Sep 26, 2026Collaborator·

  ·
  11

• Add sphereforce visualization

  #907·klei22 opened Sep 24, 2026Collaborator·

  ·
  7

• Add frozen head experiments

  #906·klei22 opened Sep 20, 2026Collaborator·

  ·
  6

• Add mesh experiment

  #904·klei22 opened Sep 4, 2026Collaborator·

  ·
  3

• feat(hangul_benchmark): Seq2Seq comparison of Hangul Factorizer vs Three-Hot Tokenizer

  #902·kahye opened Sep 2, 2026Collaborator·

  ·

• feat(korean_pos): add companion stream 256-byte fallback, OOV benchmark suite, and evaluation improvements

  #891·kahye opened Aug 16, 2026Collaborator·

  ·

• feat(korean): introduce coarse POS tag mapping for Hangul multicontext tokenizer

  #890·kahye opened Aug 15, 2026Collaborator·

  ·

• Add linear-tail squared activation variants

  #889·klei22 opened Aug 14, 2026Collaborator·

  ·
  1

• feat(attention): add relu2max residual routing

  #888·klei22 opened Aug 14, 2026Collaborator·

  ·
  4

• Dataset Expansion and Training & Loss Formulation Hangul Pos Tokenizer

  #887·kahye opened Aug 11, 2026Collaborator·

  ·

• Add input norms for mlp and attn

  #885·klei22 opened Aug 8, 2026Collaborator·

  ·
  2

• Add sweep for wqwkwowv

  #881·klei22 opened Aug 3, 2026Collaborator·

  ·
  4

• Add sweeps for input norms and wqwkt

  #880·klei22 opened Aug 3, 2026Collaborator·

  ·
  3

• feat(huggingface): add LayerNorm pruning explorer

  #879·klei22 opened Aug 3, 2026Collaborator·

  ·
  6

• Add hf semantic factorization study

  #878·klei22 opened Aug 3, 2026Collaborator·

  ·
  2

• feat(tokenizer): exclude HuggingFace input token IDs

  #877·klei22 opened Aug 3, 2026Collaborator·

  ·
  1

• Implement flexible input norms for attn and mlp

  #875·klei22 opened Aug 3, 2026Collaborator·

  ·
  3