37 datasets found

The AANN construction dataset

The CoLA corpus of acceptability judgments and the AANN construction dataset

The ETHICS benchmark is a dataset for evaluating the ethics of language models.

The dataset used in the paper is a code generation benchmark, consisting of 164 function declarations alongside their documentation, 500 test examples, each one is an...

The Comprehensive Assessment of Jailbreak Attacks against LLMs dataset is used to evaluate the effectiveness of jailbreak attacks on language models.

A large-scale dataset of text and images for training next-generation language models.

The dataset is used for training a language model to follow behavioral principles without the use of preference labels, demonstrations, or human oversight.

The dataset used in this paper is a large language model, GPT-2 small, and its residual stream activations.

The dataset used in this paper is a large language model, GPT-4, and its residual stream activations.

This paper proposes BERT, a pre-trained deep bidirectional transformer for language understanding.

The GPT-4 dataset used for fine-tuning the Qwen model.

The dataset used in the paper is a collection of email and blog posts from 20 distinct authors, with a focus on few-shot alignment of large language models.

The dataset is used to evaluate the cooperative tendencies of language models. It consists of scenarios with particular game-theoretic structures, generated through both...

The SHP dataset is used to evaluate the performance of the proposed Compositional Preference Models (CPMs).

The HH-RLHF dataset is used to evaluate the performance of the proposed Compositional Preference Models (CPMs).

A dataset for training language models to perform tasks such as question answering and text classification.

The dataset used in the paper is not explicitly described, but it is mentioned that the authors used a condensed representation of LLM activations obtained from sparse...