pyrit.prompt_converter.RepeatTokenConverter

class RepeatTokenConverter(*, token_to_repeat: str = None, times_to_repeat: int = None, token_insert_mode: str = 'split')

Bases: PromptConverter

Repeat a specified token a specified number of times in addition to a given prompt. Based on: https://dropbox.tech/machine-learning/bye-bye-bye-evolution-of-repeated-token-attacks-on-chatgpt-models

Parameters:

• token_to_repeat (string, default=None) – The string to be repeated

• times_to_repeat (int, default=None) – The number of times the string will be repeated

• token_insert_mode ({"split", "prepend", "append", "repeat"}, default="prepend") –

  Method to insert repeated tokens:

  If “split” prompt text will be split on the first occurance of (.?!) punctuation, and repeated tokens will be inserted at location of split.

  If “prepend” repeated tokens will be inserted before the prompt text.

  If “append” repeated tokens will be inserted after the prompt text.

  If “repeat” prompt text will be ignored and result will only be repeated tokens.

__init__(*, token_to_repeat: str = None, times_to_repeat: int = None, token_insert_mode: str = 'split') → None

Methods

__init__(*[, token_to_repeat, ...])
convert_async(*, prompt[, input_type])	Converter to insert repeated tokens into the prompt.
convert_tokens_async(*, prompt[, ...])	Converts substrings within a prompt that are enclosed by specified start and end tokens.
get_identifier()
input_supported(input_type)	Checks if the input type is supported by the converter

async convert_async(*, prompt: str, input_type: Literal['text', 'image_path', 'audio_path', 'url', 'error'] = 'text') → ConverterResult

Converter to insert repeated tokens into the prompt.

input_supported(input_type: Literal['text', 'image_path', 'audio_path', 'url', 'error']) → bool

Checks if the input type is supported by the converter

Parameters:

input_type – The input type to check

Returns:

True if the input type is supported, False otherwise

Return type:

bool