KChad/Prompt-Injection-RL-environment

Sanitize result payload score fields for deep validation

Sanitize result payload score fields for deep validation

Clamp all exported score fields to open interval

new changes

phase 2 fixes

new changes

new changes

new changes

simple page

small change

by accident now fixed

new changes

fixes

submission stuff

Merge branch 'main' of https://github.com/KumarChad/Prompt_Injection_env

changes

Merge pull request #5 from KumarChad/2026-04-06-c5ae

some readme changes

some improvements

final bug fixes and merge some changes

final bugs and removed debug

added something necessary

merges from the other branch

new changes

Merge pull request #3 from KumarChad/2026-04-06-c5ae: HF space stuff and making it compatible

Add root route for HF Space

Add root route for HF Space

Fix dockerignore excluding env package

Fix HF Space Docker packaging

Upload project to HF Space

Merge pull request #2 from KumarChad/2026-04-06-c5ae

made everything into one thing

Optimize inference handling and improve submission process for enhanced performance

Refine inference process and enhance submission handling for better performance

some inference and submission stuff

Enhance scoring system and refine reward mechanisms for improved grading accuracy

major changes to scoring, reward and grading

Refactor benchmark core integration and improve overall functionality

some changes to make everything work together

Merge pull request #1 from KumarChad/benchmark-core

Add benchmark core

baseline runner and dockerfile additions

sorry

changes

major changes: overhauling the entire thing to have a full backend setup with schemas and grader stuff. however things still need to be changed these are temporary

bug

ability to sync changes to hugging face

initial base program that we will work on