Leshem Choshen

Postdoctoral Researcher · MIT-IBM Watson AI Lab, IBM Research, Weizmann Institute of Science

ORCID 0000-0002-0085-6496 · leshem.choshen@weizmann.ac.il

135 papers · 5 guides · code · llms.txt

Most cited

• TIES-Merging: Resolving Interference When Merging Models 855 citations
• tinyBenchmarks: evaluating LLMs with fewer examples 276 citations
• Active Learning for BERT: An Empirical Study 244 citations
• Findings of the BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora 232 citations
• Global MMLU: Understanding and Addressing Cultural and Linguistic Biases in Multilingual Evaluation 181 citations
• An autonomous debating system 172 citations
• Q²: Evaluating Factual Consistency in Knowledge-Grounded Dialogues via Question Generation and Question Answering 167 citations
• On the Weaknesses of Reinforcement Learning for Neural Machine Translation 127 citations
• Fusing finetuned models for better pretraining 120 citations
• DisentQA: Disentangling Parametric and Contextual Knowledge with Counterfactual Question Answering 119 citations