fsid_non_clustered_0.05_threshold_trinity-nano-unlearnt
SimNPO-unlearned checkpoint of arcee-ai/Trinity-Nano-Base.
Method: SimNPO (Simple Noise-contrastive Policy Optimization)
Forget set threshold: 0.05 memorization fraction
Training: FSDP (2x A100-80GB), gradient checkpointing, 1 epoch
See unlearning-cleanslate/fsid_trinity-nano for evaluation artifacts, scores, and validation results.
- Downloads last month
- 1
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for unlearning-cleanslate/fsid_non_clustered_0.05_threshold_trinity-nano-unlearnt
Base model
arcee-ai/Trinity-Nano-Base-Pre-Anneal Finetuned
arcee-ai/Trinity-Nano-Base