File size: 930 Bytes
38ade2a
f494dbf
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
38ade2a
 
 
 
a4db835
 
38ade2a
93908f0
 
a4db835
38ade2a
a4db835
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
---
language:
- en
pipeline_tag: text-generation
tags:
- facebook
- meta
- pytorch
- llama
- llama-3
license: apache-2.0
license_name: llama3
inference:
  parameters:
    max_new_tokens: 300
    stop:
    - <|end_of_text|>
    - <|eot_id|>
---

# Model Card for Model ID

## Overview:
raaec/Phi-3-mini-4k-instruct-introvert is a language model that exhibits introverted behavior, using orthogonalization to ablate extroverted tendencies.

## !! When using the model make sure to use `tokenizer = AutoTokenizer.from_pretrained("microsft/Phi-3-mini-4k-instruct")`

## Methodology:

### Base Model: microsoft/Phi-3-medium-4k-instruct
### Orthogonalization: Applied to ablate extroverted behaviors.
### Ablation Technique: Utilizes minimal data to inhibit refusal and enhance introversion without altering other behaviors.
### Purpose:
This model is ideal for applications requiring concise, reserved responses.(sometimes a bit funny)