Instructions

• This app evaluates an LLM's ability to produce correct pilot jargon
• The brevity codes are from SD Brevity 2025 (approved for public release)
• The model is Claude Haiku 4.5 with brevity codes in its context window
• Use this app in one of two ways:
• Either provide a plain text example scenario to the model (without brevity codes), then rate the models abilty to produce a correct brevity code response
• Or provide an example scenario exclusively in brevity codes, then rate the models abilty to produce a correct text description
• Rate the models response from 1 to 5 stars and in the feedback box, explain what the correct response should be
• Do not enter CUI data
• Click Record Data to save your evaluation to the SQL database for analysis