Reveals Large Language Models' Strengths and Biases In this supplementary materials we show additional results mentioned in the main paper. First, we give experimental details in Section A. Next, we show results for Llama 2 on the bandit task in Section B. Afterwards, we show in Section C.1 additional quantitative results for the expertise-based Section D provides additional details about the vision and language tasks. For more details on the code please refer to the README.md Section A.1) and the amount of compute required to reproduce our experiments (Section Section A.2) A.1 Prompt variations generated by meta-prompting Work done whilst visiting University of Tübingen 37th Conference on Neural Information Processing Systems (NeurIPS 2023). For all Vicuna-13B based experiments (bandit, reasoning and vision) we used a single Nvidia A100-40GB GPU.

In everyday conversations, humans can take on different roles and adapt their vocabulary to their chosen roles. We explore whether LLMs can take on, that is impersonate, different roles when they generate text in-context.

Large language models have shown tremendous performance in a variety of tasks.

Training data attribution (TDA) techniques find influential training data for the model's prediction on the test data of interest.

Then, we intend to calculate the constraint part. The algorithm for Licence method for single-target interventiion scenario is shown in Algorithm 1. The details of experimental baselines are demonstrated as follows. AIT [11] is an active learning method that utilize f-score to select intervention queries. REAL fidelity means the model always choose the highest fidelity to conduct experiments.