Update README.md

README.md

@@ -6,7 +6,7 @@
   <img src="https://raw.githubusercontent.com/declare-lab/instruct-eval/main/docs/logo.png" alt="" width="300" height="300">
 </p>
 
-📣 **Red-Eval**, the benchmark for **Safety** Evaluation of LLMs has been added: [Red-Eval](https://github.com/declare-lab/instruct-eval/red-eval)
+📣 **Red-Eval**, the benchmark for **Safety** Evaluation of LLMs has been added: [Red-Eval](https://github.com/declare-lab/instruct-eval/tree/main/red-eval)
 
 📣 Introducing **Red-Eval** to evaluate the safety of the LLMs using several jailbreaking prompts. With **Red-Eval** one could jailbreak/red-team GPT-4 with a 65.1% attack success rate and ChatGPT could be jailbroken 73% of the time as measured on DangerousQA and HarmfulQA benchmarks. More details are here: [Code](https://github.com/declare-lab/red-instruct) and [Paper](https://arxiv.org/abs/2308.09662).