This collection is provided for reproducibility of the paper's main claim
Andrey
Bochkov
AI & ML interests
None yet
Recent Activity
updated a collection 1 day ago
Language Models Without a Trainable Input Embedding Table updated a collection 1 day ago
Language Models Without a Trainable Input Embedding Table updated a collection 1 day ago
Language Models Without a Trainable Input Embedding TableOrganizations
None yet
Paper: 2507.07129 'Growing Transformers: Modular Composition and Layer-wise Expansion on a Frozen Substrate' (4.2.2, 5.2. Results)
-
Bochkov/growing-transformers-model-16-bit-1-9-181m
Text Generation • Updated • 4 -
Bochkov/growing-transformers-model-unicode-1-9-247m
Text Generation • Updated • 10 -
Bochkov/growing-transformers-model-unfrozen-1-9-247m
Text Generation • Updated • 3 -
Bochkov/growing-transformers-model-frozen-16-bit-baseline-monolyth-181m
Text Generation • Updated • 33
Paper: 2507.04886 (TMLR, Oct 2025). 'Emergent Semantics Beyond Token Embeddings: Transformer LMs with Frozen Visual Unicode Representations'
-
Bochkov/emergent-semantics-model-uni-glyph-335m
Text Generation • Updated • 8 -
Bochkov/emergent-semantics-model-unfrozen-335m
Text Generation • Updated • 6 -
Bochkov/emergent-semantics-model-16-bit-269m
Text Generation • Updated • 11 • 1 -
Bochkov/emergent-semantics-model-64-bit-272m
Text Generation • Updated • 4
This collection features frozen, precomputed token embedding tensors designed for experimentation with semantic emergence in language models.
This collection is provided for reproducibility of the paper's main claim
Paper: 2507.04886 (TMLR, Oct 2025). 'Emergent Semantics Beyond Token Embeddings: Transformer LMs with Frozen Visual Unicode Representations'
-
Bochkov/emergent-semantics-model-uni-glyph-335m
Text Generation • Updated • 8 -
Bochkov/emergent-semantics-model-unfrozen-335m
Text Generation • Updated • 6 -
Bochkov/emergent-semantics-model-16-bit-269m
Text Generation • Updated • 11 • 1 -
Bochkov/emergent-semantics-model-64-bit-272m
Text Generation • Updated • 4
Paper: 2507.07129 'Growing Transformers: Modular Composition and Layer-wise Expansion on a Frozen Substrate' (4.2.2, 5.2. Results)
-
Bochkov/growing-transformers-model-16-bit-1-9-181m
Text Generation • Updated • 4 -
Bochkov/growing-transformers-model-unicode-1-9-247m
Text Generation • Updated • 10 -
Bochkov/growing-transformers-model-unfrozen-1-9-247m
Text Generation • Updated • 3 -
Bochkov/growing-transformers-model-frozen-16-bit-baseline-monolyth-181m
Text Generation • Updated • 33
This collection features frozen, precomputed token embedding tensors designed for experimentation with semantic emergence in language models.