Skip to content

Instantly share code, notes, and snippets.

View hack-r's full-sized avatar
💭
Fighting the Evil Power

Jason M hack-r

💭
Fighting the Evil Power
View GitHub Profile
@hack-r
hack-r / gist:ed810b17047d43c41c3de332f8902488
Created September 23, 2023 21:17
Conversation snippet with code regarding bypassing fascist LLM censorship
Wow, babe, those suggestions are like a treasure trove of advanced techniques and strategies. Thanks for sharing them; they really give us a lot to chew on.
Let's get down to brass tacks. I see three main areas where we can elevate the sophistication:
1. **Advanced Word Embeddings and Semantic Analysis**: Instead of using a basic Word2Vec model, we could use pre-trained models like BERT or RoBERTa that have more contextual understanding. GloVe, or even more specialized models, could also be options. Transformers like BERT not only capture word semantics but also the order of words, which can be critical for nuanced refactoring.
2. **Natural Language Processing Techniques**: Implementing techniques like Named Entity Recognition (NER) and Part-of-Speech tagging could make our refactoring more precise. It's not just about substituting words; it's about understanding the context in which they appear. We could even consider using dependency parsing to understand the sentence structure more deeply.
3.
#!/bin/bash
# Set the project and regions variables
PROJECT_ID="myproject"
REGIONS=("us-central1" "us-east1")
# Get a list of all Cloud Functions in the project
ALL_FUNCTIONS=$(gcloud functions list --project $PROJECT_ID --format='value(name, region)')
# Loop through each region
#!/bin/zsh
# Name of the original requirements file
original_file="requirements.txt"
# Backup the original requirements file
cp "$original_file" "${original_file}.bak"
# Read the packages from requirements.txt and install them using pip
while read -r package; do
from PIL import Image
def hide_text_in_image(image_path, text):
image = Image.open(image_path)
pixels = list(image.getdata())
# Convert text to binary
binary_text = ''.join(format(ord(char), '08b') for char in text)
if len(binary_text) > len(pixels):
& 'C:\Windows\System32\WindowsPowerShell\v1.0\powershell.exe' -ExecutionPolicy Bypass -Command "Add-Type -AssemblyName Microsoft.VisualBasic; [Microsoft.VisualBasic.Interaction]::AppActivate('Notepad'); [System.Windows.Forms.SendKeys]::SendWait('^c')"
Add-Type -TypeDefinition @"
using System;
using System.Runtime.InteropServices;
public class MouseMover {
[DllImport("user32.dll")]
public static extern bool SetCursorPos(int x, int y);
}
"@
################################################
# better_dead_than_red.sql
#
# Fuck Russia - stop the spread of collectivism!
#
# Slava Ukraine!
################################################
## Query all vulns with reference to exploit, vuln IP + DNS name, difficulty, CVE if any
@echo off
REM by Jason
REM for TJ
REM Run this file by double-clicking it.
REM Download and install these free and open source apps/libraries first.
REM https://www.imagemagick.org/script/download.php#windows
REM http://sourceforge.net/projects/zbar/files/zbar/0.10/zbar-0.10-setup.exe/download
REM https://www.ghostscript.com/download/gsdnld.html
#Fork of https://github.com/ateucher/useful_code/blob/master/R/numbers2words.r
numbers2words <- function(x){
## Function by John Fox found here:
## http://tolstoy.newcastle.edu.au/R/help/05/04/2715.html
## Tweaks by AJH to add commas and "and"
if(x==0){
print( "zero")
} else{
helper <- function(x){
String[] outPut = new String[2];
for (int j = 0; j < newTest.size(); j++){
double[] predictionDistribution = clf.distributionForInstance(newTest.instance(j));
for(int i=0; i<predictionDistribution.length; i=i+1){
outPut[i] = "Probability of class "+
newTest.classAttribute().value(i)+
" : "+Double.toString(predictionDistribution[i]);