HandshakeAI - Project Spectra
HandshakeAI - Project Spectra AI Trainer Assessed a range of image generations based on 5 model evaluation standards (Instruction following, Image Consistency, Quality and Naturalness and Overallness), backed by concise written justification detailing the ranking process. Each task was systematically analyzed to construct and understand the users explicit and implicit requests as well as their objective or intent. All image generations were evaluated through a rubric checklist, which specific weights or scoring systems were applied onto consistently. Took note of any generations that violated the copyright guidelines upholding privacy protocols.