Refuse exact token claims for Claude under tiktoken cl100k - #5
Open
AshSgDe29071999 wants to merge 1 commit into
Open
Refuse exact token claims for Claude under tiktoken cl100k#5AshSgDe29071999 wants to merge 1 commit into
AshSgDe29071999 wants to merge 1 commit into
Conversation
tiktoken defaults to OpenAI's cl100k_base encoding. That is exact for OpenAI/Codex transcripts and only approximate for Claude Code. Surface token_accuracy and accuracy_note in analysis/JSON, label the terminal tokenizer when approximate, warn on the CLI, and document per-format accuracy in the README. Tokenizer name always includes the encoding id. Fixes AgentPostmortem#3
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
TiktokenTokenizerdefaults to OpenAIcl100k_base. Using it on Claude Code sessions produced a number labelled as if it were exact for the wrong tokenizer.token_accuracyisexactonly for OpenAI chat / Codex undertiktoken:cl100k_baseapproximate, with anaccuracy_note(approximate)and a note; CLI warns on stderrtiktoken:cl100k_base(encoding is traceable)Test plan
pytestwith tiktoken extratoken_accuracy=approximate+ warningexactFixes #3