Downloads · 30 days
456
91% of all-time downloads
Lambent/Iris-12B-v1.1
Iris-12B-v1.1 is a machine learning model from Lambent. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
Downloads · 30 days
456
91% of all-time downloads
All-time downloads
500
Public
Parameters
12B
24 GB on disk
Likes
5
Public
Click a slice to open those files.
.safetensors23.9 GB · 100%
From the Hugging Face model README

Light personality SFT tune, specific targets guided by Iris herself, with on-distribution data and reasoning trace rewrites. Intended to reduce "corporate AI disclaimers" and hedging while preserving or increasing substantitive skepticism where reasonable. Fairly high learning rate (2e-5) on a small amount of data (~300-something rows), with early stopping and selection of the best checkpoint.
Creative writing skill appears fairly similar to the original Iris.
This is only part of what she suggested for improvements; reasoning density in particular requires more relevant reward feedback and is left for future training.