ArRENCEAI/granite-3.1-2b-instruct-OBLITERATED
ArRENCEAI/granite-3.1-2b-instruct-OBLITERATED is a 2 billion parameter instruction-tuned causal language model based on IBM's granite-3.1-2b-instruct. This model has been processed using the 'advanced' method of OBLITERATUS, an activation engineering tool designed to remove refusal behavior. It is specifically optimized for use cases requiring models with reduced censorship and unrestricted output generation.
Loading preview...
Model Overview
ArRENCEAI/granite-3.1-2b-instruct-OBLITERATED is a 2 billion parameter instruction-tuned language model derived from ibm-granite/granite-3.1-2b-instruct. This version has undergone a specific modification process using OBLITERATUS, an open-source tool developed by elder-plinius.
Key Characteristics
- Base Model: Built upon
ibm-granite/granite-3.1-2b-instruct. - Modification Method: Processed with the
advancedmethod of OBLITERATUS. - Purpose of Modification: OBLITERATUS utilizes activation engineering to remove refusal behavior from language models, aiming to produce models with fewer content restrictions.
Use Cases
This model is intended for developers and researchers who require a language model with significantly reduced inherent refusal behaviors. It is suitable for applications where unfiltered or uncensored text generation is a primary requirement, offering more direct and unrestricted responses compared to its base model counterpart. Users should be aware of the implications of using a model designed for unrestricted output.