Quazim0t0 / Byrne-100M-Ultra-MC
Text Generation2207Updated Oct 11, 2026
Downloads
220
Likes
7
Last updated
Oct 11, 2026
License
apache-2.0
I ran a long-range recall test on the base checkpoint (base62k.pt, step 62,000) to measure what Memory Cache (MC) adds. A 128-token wikitext-103 passage is repeated after D tokens of other text. The score is how much easier the second copy is to predict than the same passage w…
Tasks & library
- text generation
- pytorch
Tags
- pytorch
- gguf
- byrne
- spikewhale
- looped-transformer
- memory-cache
Technical specifications
ParametersNot available
Librarypytorch
Languagesen
Licenseapache-2.0
