Observed arrival · 2026-11-22
Latent Lens, an instrument for AI answer failures
An AI evaluation platform that tests LLM apps, RAG pipelines, and agents for grounding, hallucinations, and instruction-following, then reports failure types and fixes.
○Open
⊠Login
$Paid
†Ads
✦Pretty
●Pro
◎Niche
◉Human
⚑Risk
ƒJS
- For
- Teams evaluating LLM apps, RAG pipelines, or AI agents
- Worth noticing
- Its sample refund bot is flagged for numeric drift: a 60-day promise against a 30-day source.
Field notes
The described workflow queues evaluation jobs, decomposes responses into claims, and checks those claims against supplied ground truth. Prompt constraints are extracted and verified separately. The page says rubric and model versions are pinned so runs can be compared, with labels converted into scores in code; its example report identifies a numeric mismatch between a bot answer and its source.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
○OpenPublic substance visible
⊠LoginAccess appeared gated
✦PrettyNotable craft visible
◎NicheUnusually specific use
ƒJavaScriptBrowser-side code central
One card from the complete issue