A curated dataset for training models to distinguish between AI-generated 'slop' and quality human writing. It was created by feeding 200 prompts from ChaoticNeutrals/Reddit-SFW-Writing_Prompts_ShareGPT into various LLMs and comparing responses. The dataset was authored by DrRiceIO7 and last updated on March 24, 2026.
Use Cases
- Train text classification models based on the concept of 'slop words' versus 'quality words'
- Benchmark AI text detectors based on the comparison of LLM and human responses
- Analyze stylistic patterns in AI-generated writing based on weighted word scoring
Strengths
- Dataset is curated specifically for the task of distinguishing AI-generated 'slop' from quality writing
- Creation method involved comparing responses from multiple LLMs against human-generated responses
- Last updated on March 24, 2026
Limitations
- Column-level documentation is absent; field semantics must be inferred after download
- Row count is unknown, which may limit suitability assessment
- The dataset's scoring method ('weighted slop words') is described but its exact implementation is not detailed
Provenance
- Source
- DrRiceIO7
- Collection Method
- Created by feeding 200 prompts from ChaoticNeutrals/Reddit-SFW-Writing_Prompts_ShareGPT into various LLMs and scoring responses.
- Freshness
- Last updated 2026-03-24 16:10:19