# A New Generation of Perspective API: Efficient Multilingual Character-level Transformers

arXiv: https://arxiv.org/abs/2202.11176 · HTML: https://mlsanity.com/paper/2202.11176/

## Key facts

- Cited by 185 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 38 references that are in the map.
- First posted on arXiv 22 Feb 2022.

## Builds on (most cited first, 21 of 21)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing](https://mlsanity.com/paper/1808.06226/) (2018)
- [mT5: A massively multilingual pre-trained text-to-text transformer](https://mlsanity.com/paper/2010.11934/) (2020)
- [WILDS: A Benchmark of in-the-Wild Distribution Shifts](https://mlsanity.com/paper/2012.07421/) (2020)
- [Adafactor: Adaptive Learning Rates with Sublinear Memory Cost](https://mlsanity.com/paper/1804.04235/) (2018)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://mlsanity.com/paper/2105.13626/) (2021)
- [TweetEval: Unified Benchmark and Comparative Evaluation for Tweet Classification](https://mlsanity.com/paper/2010.12421/) (2020)
- [Nuanced Metrics for Measuring Unintended Bias with Real Data for Text Classification](https://mlsanity.com/paper/1903.04561/) (2019)
- [BERTweet: A pre-trained language model for English Tweets](https://mlsanity.com/paper/2005.10200/) (2020)
- [Ex Machina: Personal Attacks Seen at Scale](https://mlsanity.com/paper/1610.08914/) (2016)
- [Racial Bias in Hate Speech and Abusive Language Detection Datasets](https://mlsanity.com/paper/1905.12516/) (2019)
- [Mesh-TensorFlow: Deep Learning for Supercomputers](https://mlsanity.com/paper/1811.02084/) (2018)
- [Charformer: Fast Character Transformers via Gradient-based Subword Tokenization](https://mlsanity.com/paper/2106.12672/) (2021)
- [Towards generalisable hate speech detection: a review on obstacles and solutions](https://mlsanity.com/paper/2102.08886/) (2021)
- [All You Need is "Love": Evading Hate-speech Detection](https://mlsanity.com/paper/1808.09115/) (2018)
- [Hatemoji: A Test Suite and Adversarially-Generated Dataset for Benchmarking and Detecting Emoji-based Hate](https://mlsanity.com/paper/2108.05921/) (2021)
- [Overview of the HASOC track at FIRE 2020: Hate Speech and Offensive Content Identification in Indo-European Languages](https://mlsanity.com/paper/2108.05927/) (2021)
- [Towards Robust Toxic Content Classification](https://mlsanity.com/paper/1912.06872/) (2019)

## Built on by (most cited first, 30 of 185)

- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!](https://mlsanity.com/paper/2310.03693/) (2023)
- [BeaverTails: Towards Improved Safety Alignment of LLM via a Human-Preference Dataset](https://mlsanity.com/paper/2307.04657/) (2023)
- [WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs](https://mlsanity.com/paper/2406.18495/) (2024)
- [A Holistic Approach to Undesired Content Detection in the Real World](https://mlsanity.com/paper/2208.03274/) (2022)
- [UL2: Unifying Language Learning Paradigms](https://mlsanity.com/paper/2205.05131/) (2022)
- [PaLI-X: On Scaling up a Multilingual Vision and Language Model](https://mlsanity.com/paper/2305.18565/) (2023)
- [SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models](https://mlsanity.com/paper/2402.05044/) (2024)
- [PaliGemma 2: A Family of Versatile VLMs for Transfer](https://mlsanity.com/paper/2412.03555/) (2024)
- [A Pretrainer's Guide to Training Data: Measuring the Effects of Data Age, Domain Coverage, Quality, & Toxicity](https://mlsanity.com/paper/2305.13169/) (2023)
- [AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts](https://mlsanity.com/paper/2404.05993/) (2024)
- [Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails](https://mlsanity.com/paper/2501.09004/) (2025)
- [PaLI-3 Vision Language Models: Smaller, Faster, Stronger](https://mlsanity.com/paper/2310.09199/) (2023)
- [GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning](https://mlsanity.com/paper/2406.09187/) (2024)
- [GuardReasoner: Towards Reasoning-based LLM Safeguards](https://mlsanity.com/paper/2501.18492/) (2025)
- [Representation Noising: A Defence Mechanism Against Harmful Finetuning](https://mlsanity.com/paper/2405.14577/) (2024)
- [Language Model Behavior: A Comprehensive Survey](https://mlsanity.com/paper/2303.11504/) (2023)
- [Feedback Loops With Language Models Drive In-Context Reward Hacking](https://mlsanity.com/paper/2402.06627/) (2024)
- [RigorLLM: Resilient Guardrails for Large Language Models against Undesired Content](https://mlsanity.com/paper/2403.13031/) (2024)
- [Fairness in Large Language Models: A Taxonomic Survey](https://mlsanity.com/paper/2404.01349/) (2024)
- [ShieldLM: Empowering LLMs as Aligned, Customizable and Explainable Safety Detectors](https://mlsanity.com/paper/2402.16444/) (2024)
- [Towards Human-centered Proactive Conversational Agents](https://mlsanity.com/paper/2404.12670/) (2024)
- [On the Challenges of Using Black-Box APIs for Toxicity Evaluation in Research](https://mlsanity.com/paper/2304.12397/) (2023)
- [R2-Guard: Robust Reasoning Enabled LLM Guardrail via Knowledge-Enhanced Logical Reasoning](https://mlsanity.com/paper/2407.05557/) (2024)
- [Are Personalized Stochastic Parrots More Dangerous? Evaluating Persona Biases in Dialogue Systems](https://mlsanity.com/paper/2310.05280/) (2023)
- [Lost in Translation: Large Language Models in Non-English Content Analysis](https://mlsanity.com/paper/2306.07377/) (2023)
- [Gradient-Based Language Model Red Teaming](https://mlsanity.com/paper/2401.16656/) (2024)
- [LLM-jp: A Cross-organizational Project for the Research and Development of Fully Open Japanese LLMs](https://mlsanity.com/paper/2407.03963/) (2024)
- [Preference Tuning For Toxicity Mitigation Generalizes Across Languages](https://mlsanity.com/paper/2406.16235/) (2024)
- [A Review of Multi-Modal Large Language and Vision Models](https://mlsanity.com/paper/2404.01322/) (2024)

## Abstract (arXiv)

On the world wide web, toxic content detectors are a crucial line of defense against potentially hateful and offensive messages. As such, building highly effective classifiers that enable a safer internet is an important research area. Moreover, the web is a highly multilingual, cross-cultural community that develops its own lingo over time. As such, it is crucial to develop models that are effective across a diverse range of languages, usages, and styles. In this paper, we present the fundamentals behind the next version of the Perspective API from Google Jigsaw. At the heart of the approach is a single multilingual token-free Charformer model that is applicable across a range of languages, domains, and tasks. We demonstrate that by forgoing static vocabularies, we gain flexibility across a variety of settings. We additionally outline the techniques employed to make such a byte-level model efficient and feasible for productionization. Through extensive experiments on multilingual toxic comment classification benchmarks derived from real API traffic and evaluation on an array of code-switching, covert toxicity, emoji-based hate, human-readable obfuscation, distribution shift, and bias evaluation settings, we show that our proposed approach outperforms strong baselines. Finally, we present our findings from deploying this system in production.