Add LLM Proxy Babylon — inference-time optimization for low-resource …#161
Open
tverney wants to merge 1 commit intoRichardLitt:masterfrom
Open
Add LLM Proxy Babylon — inference-time optimization for low-resource …#161tverney wants to merge 1 commit intoRichardLitt:masterfrom
tverney wants to merge 1 commit intoRichardLitt:masterfrom
Conversation
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds LLM Proxy Babylon to the Software section.
LLMs allocate ~93% of training tokens to English, leaving low-resource languages with degraded reasoning, higher token costs, and weaker safety alignment. This proxy bridges that gap at inference time by selectively pre-translating prompts to English before sending to the LLM.
Real benchmarks show quality scores jumping from 0.456 to 0.949 for Thai, with 70% fewer input tokens. Research by Deng et al. (2023) shows low-resource languages have 3x the likelihood of harmful content — the translate-to-English path routes prompts through the model's strongest safety guardrails.
This complements dataset creation and model training efforts by providing an immediate, deployable solution for any existing LLM.