month-level-egocentric-recordings-20-120-days-per-participant-human-written-multiple-choice CLM-0610On month-level egocentric recordings (20-120 days per participant) with human-written multiple-choice questions, current multimodal models degrade as the question moves from inferring recurring habits, to retrieving one specific episode, to composing evidence across days — the best model reached 71.8% macro-average against a 94.2% human baseline, with several models near or below the 25% chance level on route reasoning, cross-view spatial reasoning and direction judgement.
Capability: Reasoning about time in video
Observed on
Twelve open- and closed-source MLLMs, four-option multiple-choice, 1,443 QA pairs over ~301 hours of egocentric video; no LLM-as-judge..
Sources
- Measured: per-task accuracy for 12 models plus a three-annotator human baseline (Fleiss' kappa 0.78) on a benchmark the authors built. Frame budgets differ per model by official config, so input density is not controlled; the three cognitive levels are the authors' own taxonomy, and task difficulty within a level is not independently calibrated.
Disagreeing is the most useful thing you can do here. Nothing on this page waits for a reviewer: the counts above move when evidence lands, and every source stays attached with who brought it.
Related claims
- On long-video (10-150 minute) audio-visual reasoning questions, multiple-choice accuracy overstates grounding: a large open model can answer about 55% of items correctly while scoring 8.5 on spatio-temporal evidence localization, whereas a frontier proprietary model at 73% accuracy scores 35.2 on the same evidence metric, so answer accuracy alone cannot distinguish deduction from prior-driven guessing.Reasoning about time in video · reviewed by AI
- When a document image states multi-step selection constraints in narrative text and the supporting numbers sit in several charts, multimodal models degrade steadily as either the number of logical hops (2 to 5) or the number of charts (2 to 6) grows, with the best model reaching 62.83% overall against over 90% for human annotators.Carrying a chain of inference through to its conclusion · reviewed by AI
- For multimodal LLMs, a behavioral perturbation signature (answer/confidence change under visual degradation, blank-image replacement, and grounding/relation probes) predicts which failure family an error belongs to far better than scalar confidence, yet folding the same signature into confidence as a direct penalty makes answer-versus-abstain ranking worse than scalar confidence alone.Saying it does not know · reviewed by AI
- For spoken language models answering questions over long dialogue histories, converting the history to transcripts beats keeping it as audio when the answer is a lexical detail (exact slot values across dense turns: mean 71.3% vs 42.6%), but loses when the answer depends on how something was said, where audio history wins for six of seven models and by a larger margin once lexical emotion cues are removed.Understanding speech and audio · reviewed by AI
- For 4-8B multimodal models on long-video QA, RL training on questions authored from short scene segments but answered over the whole video improves benchmark accuracy more than RL on questions authored from the whole video, because segment-authored questions target details a whole-video view overlooks; including the segment timestamp in the training instruction adds a further gain by removing referential ambiguity, and is not needed at inference.Reasoning about time in video · reviewed by AI
Notes
Drafted from the paper by a model and filed unreviewed. Visible here so it can be read, not because anyone has vouched for it: it does not move any technique's standing and does not count toward the internal scorecard. Drafted confidence: medium. Falsifier as drafted: A model that scores near human level on the cascading-reasoning tasks (route reasoning, cross-view spatial reasoning, event counting) while this ordering across the three cognitive levels disappears or reverses on the same benchmark.