From 88899e5869274f3f851899b9612bfc1ee79598fe Mon Sep 17 00:00:00 2001 From: RapidPoseidon Date: Tue, 29 Sep 2026 08:45:36 +0000 Subject: [PATCH] docs(flows): use prompt as ranking context example Co-Authored-By: Claude Opus 5.5 Co-Authored-By: lino@rapidata.ai <68745352+LinoGiger@users.noreply.github.com> --- docs/flows/ranking.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/flows/ranking.md b/docs/flows/ranking.md index 4118964086..d5cf728d09 100644 --- a/docs/flows/ranking.md +++ b/docs/flows/ranking.md @@ -59,12 +59,12 @@ flow_item = flow.create_new_flow_batch( "https://example.com/image_b.jpg", "https://example.com/image_c.jpg", ], - context="These images were generated by model X", # (1)! + context="A red fox sitting in a snowy forest at sunrise", # (1)! time_to_live=300, # (2)! ) ``` -1. Shown alongside the instruction for each comparison. +1. Shown alongside the instruction for each comparison — here, the prompt the images were generated from. 2. Stops the flow item after this many seconds and returns the responses collected so far. Up to 1 hour, and at least 60 seconds with the default flow settings; defaults to 4 minutes when omitted. ## 3. Get Results