261 citations · 743 across the 25 of their papers we have counts for
8 papers · 1 filter
Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale
Amit Roth, Ankur Samanta, Matan Halevy +2
Aligning autonomous agents with human intent remains a central challenge in modern AI. A key manifestation of this challenge is reward hacking, whereby agents appear successful und…
Language model developers should report train-test overlap
Andy K Zhang, Kevin Klyman, Yifan Mai +4
Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap which refers to the extent to which the language m…
Which transformer architecture fits my data? A vocabulary bottleneck in self-attention
Noam Wies, Yoav Levine, Daniel Jannai +1
After their successful debut in natural language processing, Transformer architectures are now becoming the de-facto standard in many domains. An obstacle for their deployment over…
PMI-Masking: Principled masking of correlated spans
Yoav Levine, Barak Lenz, Opher Lieber +4
Masking tokens uniformly at random constitutes a common flaw in the pretraining of Masked Language Models (MLMs) such as BERT. We show that such uniform masking allows an MLM to mi…
The Depth-to-Width Interplay in Self-Attention
Yoav Levine, Noam Wies, Or Sharir +2
Self-attention architectures, which are rapidly pushing the frontier in natural language processing, demonstrate a surprising depth-inefficient behavior: previous works indicate th…
On the Long-Term Memory of Deep Recurrent Networks
Yoav Levine, Or Sharir, Alon Ziv +1
A key attribute that drives the unprecedented success of modern Recurrent Neural Networks (RNNs) on learning tasks which involve sequential data, is their ability to model intricat…