datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
agent-course-final-assignment
Agent Course Final Assignment - Unified Dataset
Author: Arte(r)m Sedov
GitHub: https://github.com/arterm-sedov/
Project link: https://huggingface.co/spaces/arterm-sedov/agent-course-final-assignment
Dataset Description
This dataset is produced by the GAIA Unit 4 Agent for the Hugging Face Agents Course final assignment as part of an experimental multi-LLM agent system that demonstrates advanced AI agent capabilities. It demonstrates advanced AI agent capabilities for… See the full description on the dataset page: https://huggingface.co/datasets/arterm-sedov/agent-course-final-assignment.deductive_logical_reasoning-room_assignmentedan20-assignment2-selma-ngramsAAI_assignment2selma-ngrams-assignment2
Selma Lagerlof n-gram counts
This dataset was produced for Language Technology Assignment 2 from the
course's Swedish Selma Lagerlof corpus. Each JSONL row has an ngram string
array and its integer count. The three configurations expose unigram,
bigram, and trigram counts respectively.
The files were generated by 2-language_models.ipynb using Unicode-aware
tokenization and explicit <s> / </s> sentence markers.
Assignment_3_Dataset
Assignment 3 Dataset — QLoRA/HITL Artifacts for Green Patent Detection
Dataset Summary
This repository contains the Assignment 3 data artifacts used and produced in the advanced QLoRA workflow for green patent detection, including:
top-100 uncertainty-selected claims
QLoRA reviewed outputs
final gold labels
Part C logs/summaries required by the assignment
Transparency Note on HITL Agreement Reporting
In this Assignment 3 run, i did not manually review and… See the full description on the dataset page: https://huggingface.co/datasets/CTB2001/Assignment_3_Dataset.assignment-2-datasetsAssignment2Assignment2EDAN20_Assignment2EDAN20_Assignment_2assignment2LTAssignment2assignment3-step3Assignment4-DatasetAssignment2agentprod-assignmentassignment4-preference-datasetassignment4_step1DSAA6000Q_Assignment4_DatasetEDAN20_Assignment26000q-assignment3-part3assignment3assignment3-lima-curated-150
Assignment 3 Curated LIMA Dataset
This dataset contains the curated instruction-response pairs produced in Part 3 of Assignment 3.
Source Pipeline
Start from LIMA single-turn examples.
Use the backward model to infer instructions from responses.
Score each (generated_instruction, response) pair with Qwen/Qwen3-1.7B.
Keep examples with score >= 4.
Files
train.jsonl: curated high-quality examples for final instruction tuning
scores.jsonl: all 150 scored… See the full description on the dataset page: https://huggingface.co/datasets/sunming-giegie/assignment3-lima-curated-150.assignment4-evaluation-promptsassignment2assignment4-pairrm-preferencesassignment3-curated-lima-dataset
Assignment 3 Curated LIMA Dataset
This dataset contains the curated instruction-response pairs produced in Part 3 of Assignment 3.
Source Pipeline
Start from LIMA single-turn examples.
Use the backward model to infer instructions from responses.
Score each (generated_instruction, response) pair with Qwen/Qwen3-1.7B using few-shot prompting and a 1-5 quality rubric.
Keep examples with score >= 4.
Files
train.jsonl: curated high-quality examples for final… See the full description on the dataset page: https://huggingface.co/datasets/sunming-giegie/assignment3-curated-lima-dataset.suqi77-assignment3-curated-dataset
Assignment 3 Curated Dataset
This dataset was generated for the DSAA6000 Assignment 3 replication of the paper:
Self-Alignment with Instruction Backtranslation
Contents:
Instruction/response pairs generated from LIMA responses
Quality scores from 1 to 5
Intended use: final instruction tuning after filtering by score threshold
Main data file:
curated-qwen-150.jsonl
assignment4-pairrm-lima-qwen2p5-7b
assignment4-pairrm-lima-qwen2p5-7b
Preference dataset for DSAA6000Q Assignment 4.
Base model: Qwen/Qwen2.5-7B-Instruct
Source instructions: GAIR/lima
Pair construction: all
Number of preference pairs: 500
Files
train.jsonl: preference pairs with prompt, chosen, and rejected fields
dataset_metadata.json: run metadata