Table of Contents

Class TikTokenizerTokens

Namespace
FoundationaLLM.Common.Services.Tokenizers
Assembly
FoundationaLLM.Common.dll

Well-known special TikTokenizer token names.

public static class TikTokenizerTokens
Inheritance
TikTokenizerTokens
Inherited Members

Fields

ENDOFPROMPT

End of prompt token.

public const string ENDOFPROMPT = "<|endofprompt|>"

Field Value

string

ENDOFTEXT

End of text token.

public const string ENDOFTEXT = "<|endoftext|>"

Field Value

string

FIM_MIDDLE

Fill-In-the-Middle (FIM) middle token. For details see https://arxiv.org/abs/2207.14255 and https://github.com/openai/human-eval-infilling.

public const string FIM_MIDDLE = "<|fim_middle|>"

Field Value

string

FIM_PREFIX

Fill-In-the-Middle (FIM) prefix token. For details see https://arxiv.org/abs/2207.14255 and https://github.com/openai/human-eval-infilling.

public const string FIM_PREFIX = "<|fim_prefix|>"

Field Value

string

FIM_SUFFIX

Fill-In-the-Middle (FIM) suffix token. For details see https://arxiv.org/abs/2207.14255 and https://github.com/openai/human-eval-infilling.

public const string FIM_SUFFIX = "<|fim_suffix|>"

Field Value

string