open-unlearning/unlearn_tofu_Llama-3.2-1B-Instruct_forget10_SimNPO_lr2e-05_b4.5_a1_d1_g0.125_ep10

Hugging Face
TEXT GENERATIONConcurrent Unit Cost:1Model Size:1BQuant:BF16Context Size:32kTool Calling:SupportedPublished:May 24, 2025Architecture:Transformer Featherless Exclusive Warm

The open-unlearning/unlearn_tofu_Llama-3.2-1B-Instruct_forget10_SimNPO_lr2e-05_b4.5_a1_d1_g0.125_ep10 model is a 1 billion parameter instruction-tuned language model based on the Llama-3.2 architecture. This model is specifically designed for unlearning, utilizing the SimNPO method to forget specific information. It is optimized for scenarios requiring targeted knowledge removal while retaining general instruction-following capabilities. This model is suitable for research into model unlearning and privacy-preserving AI.

Loading preview...

Model Overview

This model, unlearn_tofu_Llama-3.2-1B-Instruct_forget10_SimNPO_lr2e-05_b4.5_a1_d1_g0.125_ep10, is a 1 billion parameter instruction-tuned language model built upon the Llama-3.2 architecture. Its primary distinguishing feature is its application of the SimNPO method for targeted unlearning, specifically designed to "forget" a subset of its training data (indicated by forget10).

Key Characteristics

  • Architecture: Llama-3.2-1B-Instruct base model.
  • Parameter Count: 1 billion parameters.
  • Unlearning Method: Utilizes the SimNPO (Simple Negative Preference Optimization) technique for selective knowledge removal.
  • Instruction-Tuned: Retains general instruction-following abilities post-unlearning.

Use Cases

This model is particularly relevant for:

  • Research in Model Unlearning: Investigating the effectiveness and mechanisms of unlearning algorithms.
  • Privacy-Preserving AI: Developing models that can remove sensitive or outdated information upon request.
  • Controlled Knowledge Management: Creating models with precisely defined knowledge boundaries.

Limitations

As indicated by the model card, specific details regarding its development, training data, evaluation, biases, risks, and environmental impact are currently marked as "More Information Needed." Users should exercise caution and conduct thorough evaluations for any specific application until further details are provided.

Finetunes
1 models