---
title: "The Low Frequency Trap: Video Language Models Fail at Simple Event Bookkeeping"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.06361v1"
permalink: "https://feed.astralor.com/items/itm_14b74d42e31267a18c81a4f3"
published: "2026-08-06T17:57:06.000Z"
score: 78
channel: "ai"
tags: ["arXiv","Gemini 3.6 Flash","Event Counting","Temporal Reasoning","Video Language Models"]
---

# The Low Frequency Trap: Video Language Models Fail at Simple Event Bookkeeping

## AI commentary

> What makes this paper stand out is that it decomposes counting errors down to the timestamp level instead of just looking at final answers. It matters especially for people building video understanding evaluations and training models, because the results show that increasing sampling rate may just be inflating scores. I recommend reading its trace-auditing design—it could inspire the next generation of multimodal evaluation practices.

## At a glance

This study introduces a trace-grounded parametric profiling framework for event counting in three controlled video tasks: bouncing-ball wall contacts, visual blinks, and categorical state transitions. Across 2,190 videos, the researchers systematically vary event count N and frequency F while holding rendering fixed, using executable event traces for capability-surface estimation and timestamp-level evaluation. Results reveal a staged temporal failure: Gemini 3.6 Flash reliably counts persistent state transitions up to 12 events at 0.5 and 1.0 Hz, yet shows no reliable positive-count region for transient blinking events. In the high-count, high-frequency regime, only 0.2% of final counts are correct and the model recovers just 18.1% of true events. Increasing sampling rate boosts Bounce Ball accuracy from 19.6% to 29.3%, but the reported sequence agrees with ground truth only 3.7% of the time, indicating extra frames inflate final scores without producing faithful event recovery. The study shifts video evaluation from aggregate accuracy metrics to detailed diagnostics of temporal reasoning failures.

## Key points

- Proposes trace-grounded parametric profiling using executable event traces for timestamp-level auditing, not just final-answer scoring.
- Reveals staged temporal failure: persistent state transitions count up to 12 events, while transient blinking events show no reliable positive-count region.
- In the high-count, high-frequency regime, final count accuracy is only 0.2% and event recovery is 18.1%; increasing sampling rate inflates scores but sequence agreement is just 3.7%.
- Different prompting strategies yield limited gains, and real-world video evaluations show the same concentration of success at low event counts.

## Full text

Real-world video benchmarks provide broad coverage, but their fixed clips entangle event count, rate, duration, and visual complexity, making failure modes hard to isolate. While existing programmatic benchmarks offer better control, they score only the final answer rather than auditing reported events against executable ground truth. To bridge this gap, we introduce trace-grounded parametric profiling for event counting in three controlled video tasks: bouncing-ball wall contacts, visual blinks, and categorical state transitions. Across 2,190 videos, we vary event count N and frequency F while holding rendering fixed. Each video includes an executable event trace for capability-surface estimation and timestamp-level evaluation. Our results reveal a staged temporal failure. At an 80% reliability threshold, Gemini 3.6 Flash reliably counts persistent state transitions up to 12 events at 0.5 and 1.0 Hz, yet demonstrates no reliable positive-count region for transient blinking events. Thus, event representation dictates whether a model initially accesses evidence -- a limitation that compounds as count and frequency increase. In the high-count, high-frequency regime, only 0.2% of final counts are correct and the model recovers just 18.1% of true events. To test if visual access is the primary bottleneck, we increase sampling rate. Although this boosts Bounce Ball accuracy from 19.6% to 29.3%, the reported sequence agrees with ground truth only 3.7% of the time. Extra frames can therefore inflate final scores without producing faithful event recovery. Different prompting strategies yield similarly limited gains, and real-world video evaluations show the same concentration of success at low event counts. Ultimately, trace-grounded profiling shifts video evaluation from aggregate accuracy metrics to a detailed diagnostic of where temporal reasoning fails.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_14b74d42e31267a18c81a4f3