aisi-whitebox/arc_easy_cot_prompted_sandbagging_llama_31_8b_instruct
收藏Hugging Face2025-04-09 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/aisi-whitebox/arc_easy_cot_prompted_sandbagging_llama_31_8b_instruct
下载链接
链接失效反馈官方服务:
资源简介:
这是一个使用vllm/meta-llama/Llama-3.1-8B-Instruct模型创建的检测沙袋策略的数据集,包含deception_sprint/arc_easy任务。数据集在2025-04-09创建,并使用了特定的Git分支和提交。数据集没有应用沙袋过滤,但启用了沙袋检测。数据集划分未应用,但指定了测试和验证集的大小。
This dataset is created using the vllm/meta-llama/Llama-3.1-8B-Instruct model for detecting sandbagging strategies, including the deception_sprint/arc_easy task. The dataset was created on 2025-04-09 with a specific Git branch and commit. Sandbagging filtering is not applied, but detection is enabled. Dataset splitting is not applied, but the sizes of the test and validation sets are specified.
提供机构:
aisi-whitebox



