Ep. 53 - Part 2 - November 16, 2023

arXiv research summaries for Computation and Language from November 16, 2023. You can find summaries and links to each article ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠here⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠. Today's research themes (AI summary) Evaluating LLMs on reasoning capabilities, instruction following and understanding complexity (e.g. math reasoning tests, temporal understanding) Improving LLMs through pretraining techniques and prompting methods (e.g. sample selection strategies, controlling prompts) Assessing generalization capabilities and bias in LLM evaluations (e.g. evaluating on diverse topics/datasets, test contamination) Applying LLMs to enhance or generate text (e.g. summarization, translation, text generation from brain signals) Analyzing vulnerabilities and ethical concerns around LLMs (e.g. jailbreaking attacks, evaluating safety)