# When in Doubt, Cascade: Towards Building Efficient and Capable Guardrails

arXiv: https://arxiv.org/abs/2407.06323 · HTML: https://mlsanity.com/paper/2407.06323/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 27 of its 37 references that are in the map.
- First posted on arXiv 8 Jul 2024.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [The Curious Case of Neural Text Degeneration](https://mlsanity.com/paper/1904.09751/) (2019)
- [Supervised Contrastive Learning](https://mlsanity.com/paper/2004.11362/) (2020)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [Language (Technology) is Power: A Critical Survey of "Bias" in NLP](https://mlsanity.com/paper/2005.14050/) (2020)
- [BBQ: A Hand-Built Bias Benchmark for Question Answering](https://mlsanity.com/paper/2110.08193/) (2021)
- [ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection](https://mlsanity.com/paper/2203.09509/) (2022)
- [BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation](https://mlsanity.com/paper/2101.11718/) (2021)
- [Nuanced Metrics for Measuring Unintended Bias with Real Data for Text Classification](https://mlsanity.com/paper/1903.04561/) (2019)
- [Classification with Valid and Adaptive Coverage](https://mlsanity.com/paper/2006.02544/) (2020)
- [Uncertainty Sets for Image Classifiers using Conformal Prediction](https://mlsanity.com/paper/2009.14193/) (2020)
- [NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails](https://mlsanity.com/paper/2310.10501/) (2023)
- [Latent Hatred: A Benchmark for Understanding Implicit Hate Speech](https://mlsanity.com/paper/2109.05322/) (2021)
- ["I'm sorry to hear that": Finding New Biases in Language Models with a Holistic Descriptor Dataset](https://mlsanity.com/paper/2205.09209/) (2022)
- [Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset](https://mlsanity.com/paper/2207.00220/) (2022)
- [Synthetic Data (Almost) from Scratch: Generalized Instruction Tuning for Language Models](https://mlsanity.com/paper/2402.13064/) (2024)
- [Human-in-the-Loop for Data Collection: a Multi-Target Counter Narrative Dataset to Fight Online Hate Speech](https://mlsanity.com/paper/2107.08720/) (2021)
- [Handling and Presenting Harmful Text in NLP Research](https://mlsanity.com/paper/2204.14256/) (2022)
- [Towards Knowledge-Grounded Counter Narrative Generation for Hate Speech](https://mlsanity.com/paper/2106.11783/) (2021)
- [The Tail Wagging the Dog: Dataset Construction Biases of Social Bias Benchmarks](https://mlsanity.com/paper/2210.10040/) (2022)
- [Beyond Denouncing Hate: Strategies for Countering Implied Biases and Stereotypes in Language](https://mlsanity.com/paper/2311.00161/) (2023)
- [NLP Systems That Can't Tell Use from Mention Censor Counterspeech, but Teaching the Distinction Helps](https://mlsanity.com/paper/2404.01651/) (2024)
- [SocialStigmaQA: A Benchmark to Uncover Stigma Amplification in Generative Language Models](https://mlsanity.com/paper/2312.07492/) (2023)
- [Counterspeakers' Perspectives: Unveiling Barriers and AI Needs in the Fight against Online Hate](https://mlsanity.com/paper/2403.00179/) (2024)
- [Detectors for Safe and Reliable LLMs: Implementations, Uses, and Limitations](https://mlsanity.com/paper/2403.06009/) (2024)
- [Neural Architecture Search for Effective Teacher-Student Knowledge Transfer in Language Models](https://mlsanity.com/paper/2303.09639/) (2023)
- [A Multi-Aspect Framework for Counter Narrative Evaluation using Large Language Models](https://mlsanity.com/paper/2402.11676/) (2024)

## Built on by (most cited first, 1 of 1)

- [A Survey on Collaborating Small and Large Language Models for Performance, Cost-effectiveness, Cloud-edge Privacy, and Trustworthiness](https://mlsanity.com/paper/2510.13890/) (2025)

## Abstract (arXiv)

Large language models (LLMs) have convincing performance in a variety of downstream tasks. However, these systems are prone to generating undesirable outputs such as harmful and biased text. In order to remedy such generations, the development of guardrail (or detector) models has gained traction. Motivated by findings from developing a detector for social bias, we adopt the notion of a use-mention distinction - which we identified as the primary source of under-performance in the preliminary versions of our social bias detector. Armed with this information, we describe a fully extensible and reproducible synthetic data generation pipeline which leverages taxonomy-driven instructions to create targeted and labeled data. Using this pipeline, we generate over 300K unique contrastive samples and provide extensive experiments to systematically evaluate performance on a suite of open source datasets. We show that our method achieves competitive performance with a fraction of the cost in compute and offers insight into iteratively developing efficient and capable guardrail models. Warning: This paper contains examples of text which are toxic, biased, and potentially harmful.