Helicone
Meta Llama Prompt Guard 2 86M
Canonical ID:
helicone/llama-prompt-guard-2-86m
86M-parameter (mDeBERTa-base) multilingual classifier that detects prompt injection and jailbreak attempts
Overview
- Provider
-
Helicone
- Family
- llama-prompt-guard-2-86m
- Release date
- 2024-10-01
- Last updated
- 2024-10-01
- Input modalities
- text
- Output modalities
- text
Pricing and Limits
Token pricing
- Input
- $0.01/1M
- Output
- $0.01/1M
- Cached input
- N/A
Context and output
- Context window
- 512
- Max output tokens
- 2
- Qualified ID
- helicone/llama-prompt-guard-2-86m
Model Summary
Meta Llama Prompt Guard 2 86M is a model listing in the Helicone provider catalog. It is categorized under the llama-prompt-guard-2-86m family.
The model supports text modalities across input/output paths. Reported context capacity is 512. Pricing is listed at $0.01/1M input and $0.01/1M output.
More From Helicone
Related models from the same provider catalog.
Helicone
Anthropic: Claude 3.5 Haiku
helicone/claude-3.5-haiku
Helicone
Anthropic: Claude 3.5 Sonnet v2
helicone/claude-3.5-sonnet-v2
Helicone
Anthropic: Claude 3.7 Sonnet
helicone/claude-3.7-sonnet
Helicone
Anthropic: Claude 3 Haiku
helicone/claude-3-haiku-20240307
Helicone
Anthropic: Claude 4.5 Haiku
helicone/claude-4.5-haiku
Helicone
Anthropic: Claude 4.5 Haiku (20251001)
helicone/claude-haiku-4-5-20251001
Helicone
Anthropic: Claude Opus 4
helicone/claude-opus-4
Helicone
Anthropic: Claude Opus 4.1
helicone/claude-opus-4-1