Industry
Introducing *HackBench* 🤡💀⏱️📊 ⏲️ 💪🔧 🔥 - Detecting & localizing reward hacking in autoresearch agents - Frontier models fail - Dataset …
Introducing *HackBench* 🤡💀⏱️📊 ⏲️ 💪🔧 🔥 - Detecting & localizing reward hacking in autoresearch agents - Frontier models fail - Dataset on HuggingFace - Recipe to collect natural & synthetic hacks Blog
Introducing HackBench 🤡💀⏱️📊 ⏲️ 💪🔧 🔥 - Detecting & localizing reward hacking in autoresearch agents - Frontier models fail - Dataset on HuggingFace - Recipe to collect natural & synthetic hacks Blog post: https://facebookresearch.github.io/RAM/blogs/hackbench 🧵1/5
Related
- The @huggingface CLI now includes a command to find the best open-source models for a given dataset! This is going to be very useful for age…
- Training models is becoming easier and easier - just look at this and TRL - especially with agents! You're missing out if you're still using…
- Thanks for sharing this. The Huggingface Hub, instead of being considered as a huge dataset, should be considered as a dynamic discovery eng…
Source: Clem Delangue (X) | 2026-10-26