What is Data poisoning?
Data poisoning is a safety or alignment concept in AI systems. It is part of the vocabulary teams use when they compare models, design prompts or review an AI system.
Definition
Data poisoning is a safety or alignment concept in AI systems. It is part of the vocabulary teams use when they compare models, design prompts or review an AI system. The term is used in safety and alignment discussions, where precision matters: two teams can use the same word to mean slightly different things. In this glossary, data poisoning is treated as a practical concept rather than a formal specification.
Why it matters
Safety terms describe how systems are constrained, monitored and evaluated for harmful or unintended behaviour. They matter because a model that performs well on benchmarks can still fail badly in an open-ended deployment.
Example
Suppose you are comparing Llama 3.3 Groq Chat with another model for a production task. Data poisoning is one of the factors behind the numbers you see: Llama 3.3 Groq Chat lists a 32,768-token context window, costs $4.06 per million input tokens and scores 87.2 on HellaSwag. Understanding the term helps you decide which of those differences actually matter for your workload.
How to evaluate it
- Check whether data poisoning is defined the same way in the documentation you are reading.
- Ask which measurement or trade-off the term is describing before comparing numbers.
- Separate vendor claims from independently observed behaviour on your own data.
- Write down the definition your team uses so reviews and evaluations stay consistent.
Related terms
Related guides
FAQ
What does Data poisoning mean?
Data poisoning is a safety or alignment concept in AI systems. It is part of the vocabulary teams use when they compare models, design prompts or review an AI system.
Why does Data poisoning matter?
Safety terms describe how systems are constrained, monitored and evaluated for harmful or unintended behaviour. They matter because a model that performs well on benchmarks can still fail badly in an open-ended deployment.
How is Data poisoning used in practice?
Suppose you are comparing Llama 3.3 Groq Chat with another model for a production task. Data poisoning is one of the factors behind the numbers you see: Llama 3.3 Groq Chat lists a 32,768-token context window, costs $4.06 per million input tokens and scores 87.2 on HellaSwag. Understanding the term helps you decide which of those differences actually matter for your workload.