Skip to page content
SAVRN
Search Plan a project

Open-weight model

manifold-1.0-1b

by Naveen Mysore nmysore/manifold-1.0-1b

manifold-1.0-1b is an open-weight model from Naveen Mysore, released under llama3.2. Its published files total 1.1 MB.

Attention manifold checkpoint for LLaMA 3.2 1B Instruct. This model extends the standard transformer attention mechanism with learned manifold structures on the value path.

Parameters—
Context—
Weights1.1 MB
Licensellama3.2
AccessOpen weights
Monthly Downloads—

Model Card

Attention manifold checkpoint for LLaMA 3.2 1B Instruct. This model extends the standard transformer attention mechanism with learned manifold structures on the value path. The checkpoint contains the manifold parameters only — the base model weights are loaded from meta-llama/Llama-3.2-1B-Instruct. See the LLM Inspector for the full visualization and editing toolkit. Paper forthcoming.

Excerpt from the card by Naveen Mysore, licensed llama3.2.

Identity and Version

Repository
nmysore/manifold-1.0-1b
Publisher
Naveen Mysore
Task
Not stated by the source
Modality
Other
Library
Not stated by the source
Parameters
Not stated by the source
Languages
Not stated by the source
Revision
03e5abc89f00a2c2d30b2ddde67317d25c038a38
First published
2026-09-28
Last updated
2026-09-28

Files and Weights

5 files, 1.1 MB in total. The weights are 2 files totalling 1.1 MB in pt.

Weights2 files · 1.1 MB
Configuration1 file · 13.5 KB
Documentation1 file · 972 B
Repository1 file · 1.5 KB
Every file
FileTypeSizeSHA-256
calibration.ptWeights562.7 KB 5691f1c39c29
formation.ptWeights560.4 KB f1a99c28b684
prototype_curve_attention.pyConfiguration13.5 KB —
README.mdDocumentation972 B —
.gitattributesRepository1.5 KB —

License and Download

License
llama3.2
Access
Open weights, no gate
Download size
1.1 MB
Download from Naveen Mysore

Released by Naveen Mysore through its official repository on Hugging Face.

Built From

Memory Requirements

PrecisionWeights in memory
As published1.1 MB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Questions About manifold-1.0-1b

What license is manifold-1.0-1b released under?

llama3.2, as its publisher declares it. Read the license text before commercial use.