Update model card metadata
Browse files
README.md
CHANGED
|
@@ -5,7 +5,7 @@ library_name: pytorch
|
|
| 5 |
tags:
|
| 6 |
- interpretability
|
| 7 |
- activation-probing
|
| 8 |
-
-
|
| 9 |
- temporal
|
| 10 |
- qwen
|
| 11 |
---
|
|
@@ -30,7 +30,9 @@ The runtime uses frozen `Qwen/Qwen3.5-0.8B` activations. The public demo path do
|
|
| 30 |
|
| 31 |
## Intended Use
|
| 32 |
|
| 33 |
-
Use these checkpoints with the TM-NLA GitHub repository to run local terminal-based semantic readouts over video windows
|
|
|
|
|
|
|
| 34 |
|
| 35 |
The checkpoints are intended for interpretability and research exploration. They are not intended as a production captioner, classifier, benchmark system, or broad video-analysis product.
|
| 36 |
|
|
@@ -42,4 +44,6 @@ Readouts can be malformed, wrong, generic, or underspecified. The activation ver
|
|
| 42 |
|
| 43 |
## Citation
|
| 44 |
|
| 45 |
-
If you publish work building on this proof of concept, cite or link the companion TM-NLA GitHub repository
|
|
|
|
|
|
|
|
|
| 5 |
tags:
|
| 6 |
- interpretability
|
| 7 |
- activation-probing
|
| 8 |
+
- vision-language-models
|
| 9 |
- temporal
|
| 10 |
- qwen
|
| 11 |
---
|
|
|
|
| 30 |
|
| 31 |
## Intended Use
|
| 32 |
|
| 33 |
+
Use these checkpoints with the TM-NLA GitHub repository to run local terminal-based semantic readouts over video windows:
|
| 34 |
+
|
| 35 |
+
https://github.com/DavWer/TM-NLA
|
| 36 |
|
| 37 |
The checkpoints are intended for interpretability and research exploration. They are not intended as a production captioner, classifier, benchmark system, or broad video-analysis product.
|
| 38 |
|
|
|
|
| 44 |
|
| 45 |
## Citation
|
| 46 |
|
| 47 |
+
If you publish work building on this proof of concept, cite or link the companion TM-NLA GitHub repository:
|
| 48 |
+
|
| 49 |
+
https://github.com/DavWer/TM-NLA
|