官方服务：
资源简介：
Multi-turn pragmatic tracking
多轮语用追踪
应用场景：
创建时间：
2026-04-13
相关数据集
henry12348/DiPlomat
DiPlomat数据集旨在评估机器在实用推理和情境对话理解方面的能力，提供了一个统一的框架来理解一般实用推理。数据集包含4,177条数据，覆盖48,900个词汇，并包含6,494个人工标注的答案，涵盖5种推理类型。数据集包含两个主要任务：实用识别和推理（PIR）以及对话问答（CQA），并提供了用于零样本NLI的数据。数据集的创建过程包括自动选择、细粒度标注和人工精炼三个步骤。
Hugging Face2023-11-13 更新550
Implicature dataset
This data set consists of conversational implicatures of utterances. Conversational implicatures are the meanings of an utterance more than what is literally stated by the utterance. The data consist
Figshare2020-08-23 更新50
DCAgent/g1_min_episodes_e1_gpt_long_sampled_swesmith_psu_d1_original_40k_glm47_traces
--- dataset_info: features: - name: conversations list: - name: content dtype: string - name: role dtype: string - name: agent dtype: string - name: model dtype
Hugging Face2026-04-16 更新40
The study of the acquisition of pragmatic inferences as an element in the choice between theoretical models in dispute
Abstract: Phenomenon known for its ambivalent behavior, presupposition has been challenging some generations of linguists regarding two complementary problems: the presupposition projection problem an
DataCite Commons2021-03-23 更新70
DCAgent/exp_rpt_stack-pytest-gpt5mini_glm_4.7_traces_jupiter
--- dataset_info: features: - name: conversations list: - name: content dtype: string - name: role dtype: string - name: agent dtype: string - name: model dtype
Hugging Face2026-03-25 更新40
