Moss image model 'ghost duplicates': single-person prompts spawn phantom copies
No_Drag8242 · reddit · 2026-10-27
Reddit user NoDrag8242 reports a generation quirk in the Moss image model: prompting for "a man standing against a white background" often produces semi-transparent "ghost" copies of the subject on both sides, and two-person prompts frequently spawn a third head peeking from behind. The post documents a typical subject-count failure mode and asks for workarounds.
More from Models
- Reddit Users Say the Ming Image Model Is Seriously Underrated — alley103 · 2026-10-27
- GPT-6.1 Terra fills form in 53.5s vs 3:21 for Sol in computer use side-by-side test — sandersted · 2026-10-27
- DeepMind staff reply mentions "Gemini 4 models" in apparent early slip — vivnat · 2026-10-27
- OpenAI math repo adds new Lean-verified papers, including an ETH refutation claim — danintheory · 2026-10-27
- OpenAI models produce Lean-verified existence proofs for QCD and Yang–Mills — danintheory · 2026-10-27
- Codex ban update: users jailbreak AI agents to bypass Daybreak filter for reverse engineering — FlolightC · 2026-10-27