Xw Q.

Key Skills

Software

No software listed

Top Subject Matter

LLM alignment and human preference data

LLM supervised training dataset preparation

LLM evaluation and rating for humanistic/social logic

Top Data Types

Text

Audio

Top Task Types

RLHF

Fine-tuning

Classification

Transcription

Freelancer Overview

AI Training Annotation Practical Work - RLHF preference annotation. Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Education includes Master of Social Work, Xinyang Normal University (2027) and Bachelor of Social Work, Xinyang Normal University (2024). AI-training focus includes data types such as Text and labeling workflows including RLHF, Fine-tuning, and Evaluation.

Labeling Experience

AI Training Annotation Practical Work - Humanistic model evaluation

Text

Performed humanistic dimension evaluation of large model outputs. Identified common-sense deviations, rigid context issues, and lack of human logic using social science/humanities knowledge. Sorted issues into structured problem lists to assist model optimization and iteration. • Commonsense deviation detection and error identification • Context/empathy and human logic gap evaluation • Issue list compilation for model improvement • Compliance with annotation standards and self-check QC

2024 - Present

AI Training Annotation Practical Work - LLM training dataset organization

TextFine-tuning

Built and organized LLM training datasets for supervised training preparation. Optimized Q&A pairs, rewrote prompt instructions, cleaned redundant or duplicate text, and unified a consistent text format. Assembled standardized training materials suitable for model fine-tuning pipelines. • Q&A pair optimization and dataset structuring • Prompt instruction writing and refinement • Text cleaning and deduplication to reduce redundancy • Training material standardization and batch preparation

2024 - Present

AI Training Annotation Practical Work - RLHF preference annotation

TextRLHF

Performed RLHF human preference ranking and evaluation for model outputs. Labeled multiple AI responses by scoring, ranking preferences, and marking unreasonable or inappropriate expressions. Produced standardized RLHF annotation data for downstream model fine-tuning and optimization. • RLHF preference scoring and ordering across multi-response candidates • Identification and marking of unreasonable expressions • Output of standardized annotation artifacts for fine-tuning • Batch annotation execution with quality inspection

2024 - Present

Survey Research Assistant - Xinyang Public Opinion Survey Projects

TextClassificationTranscription

You participated in two public opinion survey projects and were responsible for transcribing oral dialogues and conducting viewpoint classification annotation. You also handled daily discourse logic sorting and common-sense content identification to ensure materials aligned with research intent. The role required attention to detail, linguistic sensitivity, and reliable organization of qualitative interview outputs. • Transcribed oral dialogue into structured text. • Performed viewpoint classification annotation and labeling. • Sorted daily discourse logic and coherence. • Identified and checked common-sense content accuracy.

2022 - 2023

Education

Work History