Publications (7)
Two Heads are Better than One: Nested PoE for Robust Defense Against Multi-Backdoors
Victoria Graf, Qin Liu, Muhao Chen
Data poisoning backdoor attacks can cause undesirable behaviors in large language models (LLMs), and defending against them is of increasing importance. Existing defense mechanisms…
TurnWise: The Gap between Single- and Multi-turn Language Model Capabilities
Victoria Graf, Valentina Pyatkin, Nouha Dziri +2
Multi-turn conversations are a common and critical mode of language model interaction. However, current open training and evaluation data focus on single-turn settings, failing to…
Olmo 3
Team Olmo, :, Allyson Ettinger +66
We introduce Olmo 3, a family of state-of-the-art, fully-open language models at the 7B and 32B parameter scales. Olmo 3 model construction targets long-context reasoning, function…
Generalizing Verifiable Instruction Following
Valentina Pyatkin, Saumya Malik, Victoria Graf +5
A crucial factor for successful human and AI interaction is the ability of language models or chatbots to follow human instructions precisely. A common feature of instructions are…
Tulu 3: Pushing Frontiers in Open Language Model Post-Training
Nathan Lambert, Jacob Morrison, Valentina Pyatkin +20
Language model post-training is applied to refine behaviors and unlock new skills across a wide range of recent language models, but open recipes for applying these techniques lag…
C-STS: Conditional Semantic Textual Similarity
Ameet Deshpande, Carlos E. Jimenez, Howard Chen +6
Semantic textual similarity (STS), a cornerstone task in NLP, measures the degree of similarity between a pair of sentences, and has broad application in fields such as information…
Pretraining Data Can Be Poisoned through Computational Propaganda
Victoria Graf, Hannaneh Hajishirzi, Noah A. Smith +2
The paper shows that language model pretraining data can be poisoned through publicly editable web discussion pages, and introduces a method called HalfLife to estimate how much ma…