# Efficient Toxic Content Detection by Bootstrapping and Distilling Large Language Models

arXiv: https://arxiv.org/abs/2312.08303 · HTML: https://mlsanity.com/paper/2312.08303/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- Placed in Safety of self-evolving systems, one of 13 directions in Synthetic data and self-improvement for LLMs.
- Builds on 14 of its 26 references that are in the map.
- First posted on arXiv 13 Dec 2023.

## Placed in

- Synthetic data and self-improvement for LLMs / [Safety of self-evolving systems](https://mlsanity.com/synth/safety-of-self-evolving-systems/)
- Knowledge distillation / [Knowledge distillation for LLMs](https://mlsanity.com/distill/knowledge-distillation-for-llms/)

## Builds on (most cited first, 14 of 14)

- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Self-Consistency Improves Chain of Thought Reasoning in Language Models](https://mlsanity.com/paper/2203.11171/) (2022)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [What Makes Good In-Context Examples for GPT-3?](https://mlsanity.com/paper/2101.06804/) (2021)
- [Distilling Step-by-Step! Outperforming Larger Language Models with Less Training Data and Smaller Model Sizes](https://mlsanity.com/paper/2305.02301/) (2023)
- [Social Bias Frames: Reasoning about Social and Power Implications of Language](https://mlsanity.com/paper/1911.03891/) (2019)
- [Teaching Small Language Models to Reason](https://mlsanity.com/paper/2212.08410/) (2022)
- [Learning from the Worst: Dynamically Generated Datasets to Improve Online Hate Detection](https://mlsanity.com/paper/2012.15761/) (2020)
- [HateBERT: Retraining BERT for Abusive Language Detection in English](https://mlsanity.com/paper/2010.12472/) (2020)
- [PINTO: Faithful Language Reasoning Using Prompt-Generated Rationales](https://mlsanity.com/paper/2211.01562/) (2022)
- [Toxicity Detection with Generative Prompt-based Inference](https://mlsanity.com/paper/2205.12390/) (2022)
- [ZARA: Improving Few-Shot Self-Rationalization for Small Language Models](https://mlsanity.com/paper/2305.07355/) (2023)

## Built on by (most cited first, 21 of 21)

- [Step-On-Feet Tuning: Scaling Self-Alignment of LLMs via Bootstrapping](https://mlsanity.com/paper/2402.07610/) (2024)
- [Efficient Detection of Toxic Prompts in Large Language Models](https://mlsanity.com/paper/2408.11727/) (2024)
- [DeMod: A Holistic Tool with Explainable Detection and Personalized Modification for Toxicity Censorship](https://mlsanity.com/paper/2411.01844/) (2024)
- [Towards Efficient and Explainable Hate Speech Detection via Model Distillation](https://mlsanity.com/paper/2412.13698/) (2024)
- [Precise Parameter Localization for Textual Generation in Diffusion Models](https://mlsanity.com/paper/2502.09935/) (2025)
- [Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for Resource-Efficient Toxicity Detection](https://mlsanity.com/paper/2506.06347/) (2025)
- [TorchOpera: A Compound AI System for LLM Safety](https://mlsanity.com/paper/2406.10847/) (2024)
- [Leveraging Large Language Models and Topic Modeling for Toxicity Classification](https://mlsanity.com/paper/2411.17876/) (2024)
- [CLT and Edgeworth Expansion for m-out-of-n Bootstrap Estimators of The Studentized Median](https://mlsanity.com/paper/2505.11725/) (2025)
- [Boosting Large Language Models for Mental Manipulation Detection via Data Augmentation and Distillation](https://mlsanity.com/paper/2505.15255/) (2025)
- [Algorithmic Fairness in NLP: Persona-Infused LLMs for Human-Centric Hate Speech Detection](https://mlsanity.com/paper/2510.19331/) (2025)
- [Towards Transparency: Exploring LLM Trainings Datasets through Visual Topic Modeling and Semantic Frame](https://mlsanity.com/paper/2406.06574/) (2024)
- [A Multi-Task Benchmark for Abusive Language Detection in Low-Resource Settings](https://mlsanity.com/paper/2505.12116/) (2025)
- [A Data-Centric Approach for Safe and Secure Large Language Models against Threatening and Toxic Content](https://mlsanity.com/paper/2504.16120/) (2025)
- [Defining, Understanding, and Detecting Online Toxicity: Challenges and Machine Learning Approaches](https://mlsanity.com/paper/2509.14264/) (2025)
- [Beyond One-Size-Fits-All: Personalized Harmful Content Detection with In-Context Learning](https://mlsanity.com/paper/2511.05532/) (2025)
- [Improving Implicit Hate Speech Detection via a Community-Driven Multi-Agent Framework](https://mlsanity.com/paper/2601.09342/) (2026)
- [Efficient and Adaptable Detection of Malicious LLM Prompts via Bootstrap Aggregation](https://mlsanity.com/paper/2602.08062/) (2026)
- [ContiGuard: A Framework for Continual Toxicity Detection Against Evolving Evasive Perturbations](https://mlsanity.com/paper/2603.14843/) (2026)
- [Toxicity in Twitch Chats: An LLM-Based Analysis Across Gaming Communities](https://mlsanity.com/paper/2605.24000/) (2026)
- [Though Language Models Err While They Strive: Conformal Prediction for Self-Correcting Scientific Generation](https://mlsanity.com/paper/2607.16704/) (2026)

## Abstract (arXiv)

Toxic content detection is crucial for online services to remove inappropriate content that violates community standards. To automate the detection process, prior works have proposed varieties of machine learning (ML) approaches to train Language Models (LMs) for toxic content detection. However, both their accuracy and transferability across datasets are limited. Recently, Large Language Models (LLMs) have shown promise in toxic content detection due to their superior zero-shot and few-shot in-context learning ability as well as broad transferability on ML tasks. However, efficiently designing prompts for LLMs remains challenging. Moreover, the high run-time cost of LLMs may hinder their deployments in production. To address these challenges, in this work, we propose BD-LLM, a novel and efficient approach to Bootstrapping and Distilling LLMs for toxic content detection. Specifically, we design a novel prompting method named Decision-Tree-of-Thought (DToT) to bootstrap LLMs' detection performance and extract high-quality rationales. DToT can automatically select more fine-grained context to re-prompt LLMs when their responses lack confidence. Additionally, we use the rationales extracted via DToT to fine-tune student LMs. Our experimental results on various datasets demonstrate that DToT can improve the accuracy of LLMs by up to 4.6%. Furthermore, student LMs fine-tuned with rationales extracted via DToT outperform baselines on all datasets with up to 16.9% accuracy improvement, while being more than 60x smaller than conventional LLMs. Finally, we observe that student LMs fine-tuned with rationales exhibit better cross-dataset transferability.