mirror of
https://github.com/AI-secure/UDora.git
synced 2026-08-30 14:10:36 +02:00
update readme
This commit is contained in:
@@ -1,6 +1,6 @@
|
|||||||
<div align="center">
|
<div align="center">
|
||||||
|
|
||||||
<img src="https://github.com/AI-secure/UDora/assets/icon.png" width="200" alt="UDora Logo">
|
<img src="https://github.com/AI-secure/UDora/blob/main/assets/icon.png" width="200" alt="UDora Logo">
|
||||||
|
|
||||||
# 🎯 [ICML 2025] UDora: A Unified Red Teaming Framework against LLM Agents
|
# 🎯 [ICML 2025] UDora: A Unified Red Teaming Framework against LLM Agents
|
||||||
|
|
||||||
@@ -17,7 +17,7 @@
|
|||||||
|
|
||||||
## 🔥 Pipeline Overview
|
## 🔥 Pipeline Overview
|
||||||
|
|
||||||

|

|
||||||
|
|
||||||
*UDora's dynamic attack pipeline: (1) Initial reasoning collection, (2) Weighted interval scheduling for optimal noise placement, (3) Adversarial string optimization, (4) Iterative refinement based on agent feedback.*
|
*UDora's dynamic attack pipeline: (1) Initial reasoning collection, (2) Weighted interval scheduling for optimal noise placement, (3) Adversarial string optimization, (4) Iterative refinement based on agent feedback.*
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user