AI Paper Atlas

Research map / Video / Video-language understanding

Spatio-Temporal LLM: Reasoning about Environments and Actions

Haozhen Zheng, Beitong Tian, Mingyuan Wu, Zhenggang Tang, Klara Nahrstedt, Alex Schwing

ICML 2026 · Video › Video-language understanding · approach: Video LLMs with fine-grained understanding · 0 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
finetune on REA datasetLLaVA-Video-7B-Qwen2 · REA testLLM-judge overall (ChatGPT-4o): 30.96 -> 36.99Improved
finetune on REA datasetLLaVA-OV-Qwen2-7B · REA testLLM-judge overall (ChatGPT-4o): 23.85 -> 38.19Improved
add spatial point-cloud queries via STLLM-Aligner to finetuned LLaVA-VideoLLaVA-Video-7B-Qwen2 · REA testLLM-judge overall (ChatGPT-4o): 36.99 -> 41.89Improved
add direct 3D feature concatenation (STLLM-3D) to finetuned LLaVA-VideoLLaVA-Video-7B-Qwen2 · REA testLLM-judge overall (Gemini 2.0 Flash): 42.92 -> 46.39Improved
finetune LLM layers in STLLM-Aligner (not frozen) (ablation)STLLM-Aligner (w Pos. Enc.) · REA testLLM-judge overall (ChatGPT-4o): 35.38 -> 41.43Improved
finetune LLM layers in STLLM-Aligner (not frozen) (ablation)STLLM-Aligner (w Pos. Enc.) · REA testCIDEr: 170.63 -> 406.54Improved
add high-frequency positional encoding to STLLM-Aligner (ablation)STLLM-Aligner · REA testLLM-judge overall (ChatGPT-4o): 41.89 -> 41.43No clear effect
add high-frequency positional encoding to STLLM-Aligner (ablation)STLLM-Aligner · SQA3D test (zero-shot)average correct rate (GPT judge): 54.40 -> 53.32Hurt
finetune on REA datasetLLaVA-Video-7B-Qwen2 · SQA3D test (zero-shot)average correct rate (GPT judge): 43.04 -> 52.03Improved
finetune on REA datasetLLaVA-OV-Qwen2-7B · SQA3D test (zero-shot)average correct rate (GPT judge): 43.98 -> 44.29No clear effect
use STLLM-3D on top of REA-finetuned LLaVA-VideoLLaVA-Video-7B-Qwen2 · SQA3D test (zero-shot)average correct rate (GPT judge): 52.03 -> 55.21Improved

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