AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Mass-Producing Failures of Multimodal Systems with Language Models

Shengbang Tong, Erik Jones, Jacob Steinhardt

NeurIPS 2023 · Vision-language and multimodal › Human and embodied modeling · approach: Multimodal and affective interaction modeling · 6 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace CLIP-similarity scraping with random COCO pairs (ablation)MultiMon · downstream text-to-image failure ratefailure rate: 80% -> 20%Hurt
remove scraped individual failures from categorization prompt (ablation)MultiMon (GPT-4) · CLIP systematic failuressystematic failures found: 14 -> 2Hurt
remove scraped individual failures from categorization prompt (ablation)MultiMon (GPT-4) · CLIP systematic failuresaverage success rate: 29.3%Hurt
use GPT-4 instead of GPT-3.5 as categorizer (ablation)MultiMon · CLIP systematic failuressystematic failures found: 8 -> 14Improved
use Claude instead of GPT-4 as categorizer (ablation)MultiMon · CLIP systematic failuressystematic failures found: 14 -> 11Hurt
use GPT-3.5 instead of GPT-4 as categorizer (ablation)MultiMon · shared systematic failuresaverage success rate: 80.2% -> 69.5%Hurt
use Claude instead of GPT-4 as categorizer (ablation)MultiMon · shared systematic failuresaverage success rate: 80.2% -> 83.3%No clear effect
use Claude or GPT-3.5 as generator instead of GPT-4 (ablation)MultiMon · CLIP success ratesuccess rateMixed
use COCO vs SNLI as input corpus (ablation)MultiMon · systematic failures foundfailure typesMixed
steer scraping toward self-drivingMultiMon · self-driving systematic failuressuccess rate: 100%Improved
steer scraping toward self-driving: object-specific attributes failureMultiMon · self-driving systematic failuressuccess rate: 41.0%Hurt
steer generation toward self-drivingMultiMon · self-driving individual failuressuccess rate: 74.56%Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy