- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
TASA proposes token-adaptive activation steering to improve video temporal grounding without updating the Video-LLM backbone. It extracts local correction directions from negative gradients of the timestamp prediction loss. Spherical K-means compresses these directions into a compact steering-vector bank. A token-level Router dynamically combines the vectors based on the current hidden state. A Gate controls the intervention strength during autoregressive decoding. The method uses three-stage training to initialize and refine the vector bank and controller. Experiments across five Video-LLMs and three benchmarks show consistent improvements, including a 12.35-point mIoU gain on Charades-STA with Qwen3-VL-8B.
Critical strengths
-
Consistent performance improvements. TASA improves temporal grounding across all 15 model–dataset settings, including cross-dataset evaluation on QVHighlights.
-
Parameter-efficient adaptation. The method keeps the backbone frozen and optimizes only a small intervention module with approximately 35K–70K trainable parameters.
-
Useful empirical analysis. Ablations examine vector initialization, bank size, and intervention depth, providing evidence for the effectiveness of gradient-derived multi-vector steering.
Critical weaknesses
-
Insufficient comparison with alternative steering methods. Gradient-derived directions are compared mainly against random initialization, without matched contrastive-steering baselines. The gains over single-layer LoRA are also marginal (0.14–0.74 mIoU).
-
Unclear consistency and roles of steering directions. The paper lacks analysis of vector stability across runs and their correspondence to specific boundary errors, such as early/late starts or ends.
-
The paper does not distinguish improvements from correcting temporal boundaries of correctly identified events versus recovering from incorrect event localization.
-
Incomplete component ablations. The individual contributions of the Gate and regularization losses remain unclear. The necessity of the three-stage optimization procedure is also insufficiently validated.
-
Limited generalization evaluation. More challenging multi-segment grounding (OMTG-Bench) and general video understanding benchmarks (Video-MME, Video-MME-v2) are missing.
-
Insufficient efficiency analysis. The paper reports trainable parameter counts but lacks end-to-end training time, GPU memory, and inference latency comparisons with LoRA.
Questions for authors
See Weaknesses.
The authors have not responded. All three reviewers find that TASA performs roughly on par with single-layer LoRA and question what it offers beyond existing adapters. My concern about the Base evaluation protocol, where Base scores substantially below the official Qwen3-VL results, also remains unaddressed. I maintain my rating of 2.