2b-8b-video-multimodal-models-fine-tuned-temporal-reasoning-placing-small CLM-0526For 2B-8B video multimodal models fine-tuned for temporal reasoning, placing a small minority of reversed-video rollouts in each GRPO group (1 reversed out of 7) beats both all-forward GRPO and larger reversed shares (2 or 3 of 7), because more reversed samples dilute the main forward-video answering objective.
Capability: Reasoning about time in video
Observed on
Qwen3-VL 2B/4B/8B and InternVL2.5-8B, action and temporal video QA benchmarks, after a supervised temporal-reasoning distillation stage; group size fixed at 7.
Sources
- supportsCOMET: Contrastive Motion-Enhanced Temporal Reasoning for Video Multimodal Large Language ModelsMeasured, single-paper for the ratio sweep but consistent with a prior independent result; the sweep is on one backbone (8B) and one group size, benchmarks are 1,000-item random samples, and no variance or seed information is reported, so the 0.4-1.2 point differences between ratios are not shown to exceed noise.
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
- For 2B-8B video multimodal models, adding a small fraction of time-reversed video rollouts into each GRPO group (e.g. 1 reversed out of 7) raises accuracy on action and temporal-reasoning video QA over standard GRPO, while raising the reversed share further (2 or 3 of 7) gives back the gain.Reasoning about time in video · 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
- In GRPO-style RL post-training of 7B multimodal reasoning models, composing each optimization group from rollouts stratified by length-normalized sequence surprisal and outcome (confident-correct anchors, confident-wrong negatives, low-confidence successes) within the same step keeps path entropy from collapsing and avoids the late-training performance decline GRPO and DAPO show after a few hundred steps.Retaining skills and safety after fine-tuning · reviewed by AI
- For deepfake video detection with 7-8B open multimodal backbones, splitting the analysis into four separately prompted perspective agents (texture, lighting, motion, physics) whose reports a judge agent reconciles beats the same backbone doing chain-of-thought or multi-turn self-questioning over the same four perspectives, and the gain is not explained by extra turns or repeated frame access.Getting more from several models than from one · reviewed by AI
- For 8B multimodal models, preference tuning and plain outcome-reward GRPO lower hallucination rates while making stated confidence worse calibrated than the untuned backbone, because rank- or correctness-based rewards never score the numerical accuracy of the model's own confidence; adding a Brier-plus-group-ECE calibration term to the GRPO reward cuts calibration error sharply and leaves accuracy metrics largely unchanged.Grounding answers in the image · 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: An ablation at matched group size and starting checkpoint where a higher reversed share (2 or 3 of 7) matches or beats 1 of 7 on action and temporal video QA averages. Drafted stance toward 2b-8b-video-multimodal-models-adding-small-fraction-time-reversed-video: supports -- An independent architecture and training pipeline reproduces the same ordering: one reversed rollout out of seven is best and the gain shrinks as the reversed share grows to 2 or 3.