Synthetic efficiency (HELM)

HELM runtime probe: generate from fixed public-domain prompts while varying prompt length, output length and tokenizer.

Also known as: HELM synthetic efficiency

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categorygeneration
Subcategorycontrolled inference-runtime probe over prompt length, output length and tokenizer
Page statusunknown
Metricinference runtime (denoised and idealized); run spec also attaches exact_match
Directionlower_is_better
Dataset size10
PublisherStanford CRFM (HELM)

What it measures

synthetic_efficiency is not a quality ranking. HELM feeds short public-domain prompts of a chosen token length and asks the model to generate a chosen number of tokens so that runtime can be compared across providers. The scenario docstring lists four questions: how runtime scales with prompt and output length and with the number of completions; how much variance each query shows; how providers differ; and whether hardware can be reverse-engineered. Prompts are stored on a CodaLab bundle and sliced per tokenizer. English-script public-domain text used as filler, not as a comprehension test.

Task format

Completion generation. Run spec synthetic_efficiency with num_prompt_tokens, num_output_tokens, tokenizer and optional random. Each run loads 10 instances. References are empty strings tagged correct. HELM also attaches exact_match and generative-harms metrics even though the gold continuation is empty.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub