- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
CAP-MED is a protocol for finding sets of internal activations in a video diffusion transformer that support a specific motion effect. For a scene prompt and seed, an instructed rollout (e.g. "pan right") and an empty-action rollout are generated; the motion effect is the change in a VideoMAE classifier logit difference. A node is the full token-by-channel output of a submodule (SA/FFN/CA) at one layer and one denoising step. EAP-IG edge attribution is summed per layer, and layers are ranked by target-vs-other specificity (target attribution divided by the mean attribution to the other nine signed directions); the top-6 layers x steps {0,2,4} x {SA, FFN} give a 36-node set. Sufficiency inserts 's activations into the rollout at those nodes and measures recovery of the effect; necessity does the reverse. On Wan2.2 T2V-A14B (10 steps, 40 layers) over ten signed camera/object directions, specificity-selected sets reach median normalized recovery/suppression 0.99/0.97 versus 0.15/0.59 for the top-36 nodes by raw attribution magnitude; same-grid random-layer and high-attribution/low-specificity controls and a linear-probe selector do worse. Independent readouts (VBench CameraPredict, CoTracker3, dense flow) preserve the ordering. Head-level experiments show a two-set composed dependence and that exchanging V (not QK) in pan heads carries the effect. On LTX-Video the workflow is rerun with a per-node magnitude-balanced variant for camera motion and color temperature, with mixed per-direction results.
Critical strengths
1. An extensive and well-controlled intervention protocol. Matched-pair bidirectional patching, same-grid controls that separate layer choice from allocation, a probe-based selector, readouts not used for selection, reference-based and reference-free quality metrics, paired bootstrap intervals, budget/step/layer sweeps, and a direct CA-addition test. Mechanistic analysis of video DiTs is underexplored and the protocol is reusable; the appendix is unusually complete.
2. Scope is stated carefully. The paper repeatedly distinguishes what the exchanges show (functional dependence of a measured score) from what they do not (a minimal circuit, content-preserving editing).
Critical weaknesses
1. The intervention unit is so coarse that near-complete recovery is expected, and the main baseline is weak. A "node" is the entire output field of SA or FFN at a layer and step; the selected set transplants 6 layers x 2 submodules at 3 of the 10 steps, i.e. a large slice of the early residual-stream computation, including appearance (LPIPS 0.51 from patched to the empty-action video; acknowledged in App. E). At that granularity is not surprising: Table 20 shows that any three late layers (37 to 39) at those steps already give , that step-0 nodes are useless () and step 2 alone gives 0.99. The headline contrast is with "EAP-IG top-36 individual nodes", whose failure is largely an allocation artefact (it spreads nodes over steps and layers where patching does nothing). The same-grid controls show that layer choice on the grid matters, but not that the specificity ratio is the right criterion: the obvious heuristics are never compared (last-k layers; top-k by target magnitude on the same grid; a step sweep). What is learned about the model, beyond "camera motion is set in mid/late layers at steps 2 to 4 and carried by V in pan heads", is thin for an interpretability paper, and the authors disclaim circuit-level conclusions.
2. The central claim (specificity beats magnitude) does not replicate cleanly outside Wan2.2 camera motion. On LTX-Video the selector had to be changed to a per-node magnitude-balanced score, raw EAP-IG has larger median effects in both pan directions and both arms (Table 25), sign-correctness orderings vary by cell, and the pooled numbers carry no uncertainty. On Wan2.2, object-lateral specificity ratios are below one and effects are weak until the budget is doubled, and object-depth sufficiency is partial. The abstract's numbers therefore describe a Wan2.2 camera-motion result obtained with a selector whose grid (steps {0,2,4}, SA/FFN, K=6) was chosen on that model; generality is asserted more than shown.
Presentation note. Many tables, several distinct selected sets (36-node grid, K=100 head sets, top-300 head set, LTX per-node sets) and dense caveats obscure the two or three takeaways. Section 4.7 opens with a sentence lacking a subject ("Instead of the Wan layer-grid rule, it ranks...").
Questions for authors
Q1. At the same 36-node budget and the same step/submodule grid, how do these compare with specificity ranking: the last six layers; the six layers with highest target-attribution magnitude; six layers chosen by a leave-one-out patching sweep? If a trivial heuristic matches specificity, the contribution reduces to the protocol.
Q2. How much of the recovery is appearance transfer? Please report recovery as a function of content change (e.g. patch a random fraction of tokens or channels within the selected fields, or report per unit LPIPS), or restrict the exchange to a motion-relevant subspace.
Q3. Please add paired uncertainty for the LTX full-workflow comparison (Tables 7 and 25) and repeat attribution with different discovery pairs. Is the specificity advantage statistically supported anywhere on LTX?
Q4. In Eq. (4) the "other" set includes the opposite direction on the same axis, whose attribution to is sign-correlated with the target's. How do the selected layers and results change if the opposite direction is excluded from the denominator?
Q5. The classifier readouts were trained on Wan2.2 generations annotated by reviewers who saw the instruction. Dense-flow directional agreement on unpatched instructed videos is only 62 to 68% for object motion. For object directions, how should the reader interpret when the ground-truth motion itself is ambiguous in a third of cases?
5 comments
Official Comment for Reviewer H8zp (2/4)
Q2. Motion effect per perceptual change and reference fidelity
We thank the reviewer for this question. We answer it in four ways, including the effect per unit LPIPS that the reviewer suggested.
What can transfer. Each pair shares the scene prompt, seed and sampler and differs only in the appended instruction. A full-field exchange therefore cannot import unrelated appearance. It can transfer only differences caused by the instruction, which include the motion itself and its accompanying changes, such as newly revealed regions under camera motion.
The patched change tracks the instructed change.
| Pooled sufficiency (App. E) | CAP-MED patched | Instructed video | EAP-IG patched |
|---|---|---|---|
| Subject consistency change | |||
| Background consistency change | |||
| Aesthetic quality change | |||
| FVD to instructed reference |
Changes are relative to the empty-action video. CAP-MED changes the video by about as much as the instruction itself does, and its outputs are closer to the instructed videos than EAP-IG outputs are. EAP-IG changes the video less because it also induces much less motion (median ). The patched-to-empty LPIPS of is therefore consistent with the instruction's own effect on the frames, since camera motion itself changes frame content.
Effect per unit LPIPS. On Pan+, the paired median gain of SPEC in is in sufficiency and in necessity over same-grid magnitude, and and over global top-36.
Comparison with the last six layers. Sufficiency, forty original pairs per direction, mean (median).
| Direction | DINO SPEC | DINO last-six | LPIPS SPEC | LPIPS last-six |
|---|---|---|---|---|
| Pan+ | ||||
| Pan- | ||||
| Tilt+ | ||||
| Tilt- | ||||
| Zoom+ | ||||
| Zoom- | ||||
| Lateral+ | ||||
| Lateral- | ||||
| Depth+ | ||||
| Depth- |
Reference is the empty-action video. DINO is the mean aligned DINOv1 B/16 cosine over 81 frames, and LPIPS uses frames . In the camera directions, where specificity reaches near-complete effects (Table 1), specificity alters reference content less than the last six layers for tilt and zoom on both DINO and LPIPS, and the two are similar for pan. Effective layer sets therefore differ in how much content they move, and specificity selects sets that move less.
Restricting the exchange. Within selected heads, exchanging only the value operand under the recipient's routing retains the effect (), whereas exchanging query–key routing does not (, Table 6).
In the revised paper, we will move these reference comparisons from App. E into the main text and report effect together with content change.
Official Comment for Reviewer H8zp (3/4)
W2 + Q3. Generality and paired LTX uncertainty
We thank the reviewer for this careful assessment and for requesting paired uncertainty, which we now report.
Wan2.2 object motion. Object effects are partial for lateral+ and depth+, but the specificity advantage holds there as well. In sufficiency, specificity exceeds EAP-IG in all four object directions (median vs for lateral+, vs for depth+, vs for lateral−, vs for depth−, Table 1). Camera-compensated CoTracker3, which is not used for selection, gives the same ordering, with sign-correctness of vs in the negative-direction sufficiency cells () and vs across object necessity (). The step and budget choices are reported with their sweeps in App. F.
LTX selector. LTX ranks individual nodes rather than layers. At the node level, a pure ratio can favor nodes whose attribution is small in every direction. LTX therefore uses , the geometric mean of target magnitude and specificity, with no fitted weight. Both selectors share the same 840 candidates and the same 36-node budget.
Q3. Is the advantage statistically supported on LTX? Yes. Paired 95% intervals exclude zero in favor of the specificity-based selector for Tilt+ sufficiency, Zoom+ in both exchanges and Zoom− necessity. Entries are median paired MB-minus-Raw normalized-flow differences (discovery , attribution seed 1).
| Direction; evaluation | Sufficiency | Necessity |
|---|---|---|
| Pan+ | ||
| Pan- | ||
| Tilt+ | ||
| Tilt- | ||
| Zoom+ | ||
| Zoom- |
The Zoom+ gain is positive in both exchanges across all five archived attribution seeds. Intervals resample matched scenes with the selected sets fixed.
Repeated discovery. We repeated attribution with two disjoint sets of twenty discovery scenes. Pan+, which shows no significant difference in the original setting, shows a significant gain on one set (sufficiency , necessity ), while the other set's intervals include zero. Both repetitions use the same forty evaluation pairs. Selection is also stable to the discovery sample. With calibration fixed, omitting one discovery scene retains of nodes on average across six directions.
In the revised paper, we will add these intervals to Section 4.7, state that LTX supports the advantage for tilt and zoom, and make clear in the abstract that the reported numbers are from Wan2.2.
Q4. Excluding the opposite direction
We thank the reviewer for this suggestion. Excluding the opposite direction leaves the selection and the effect essentially unchanged.
| Opposite-exclusion check | Result |
|---|---|
| Retained layers; ten directions | |
| Changed Pan+ layer | |
| Modified Pan+ ; , sufficiency/necessity | |
| Sign agreement; both exchanges | |
| Paired change versus original; sufficiency/necessity |
Across the ten directions, 52 of the 60 selected layers are retained. For Pan+, only one layer changes (27 to 0), and the modified set gives the same effect as the original set, with paired changes below and sign agreement in every pair.
The opposite direction's mass need not mirror the target's. Although , is computed from the opposite instruction's own prompt pair, so its activation difference and attribution mass differ from those of the target. Including it in the denominator penalizes layers that are attributed to the whole axis regardless of sign, which favors direction-specific layers.
In the revised paper, we will report this opposite-excluded variant in the appendix.
Official Comment for Reviewer H8zp (1/4)
We thank the reviewer for the thorough and constructive review, and for recognizing the extensive and well-controlled intervention protocol, its reusability, the completeness of the appendix and our careful statement of scope. The suggested controls were especially valuable. Following them, we report the requested same-grid heuristics, content measurements across all ten directions, paired uncertainty and repeated discovery on LTX-Video, and the opposite-exclusion check. We address each point below and hope that these responses resolve the reviewer's concerns.
W1 + Q1. Identical-grid controls and scientific contribution
We thank the reviewer for this detailed critique and for proposing these controls. We ran all three on the identical grid.
All selectors share the original pairs and the six-layer {0,2,4} SA/FFN grid. Sufficiency inserts instructed activations into the empty-action run, and necessity exchanges in the opposite direction. The leave-one-out sweep used 1,680 forward passes, 1,640 for selection and 40 for endpoints.
Coarse units do not make recovery automatic. At the same 36-field budget and grid, the six layers with the largest target attribution give median ρ of −0.02 and 0.02. Fixed three-layer sets at the same steps give −0.02 for layers 15 to 17 and −0.06 for layers 5, 22 and 35 (Table 20). Near-complete recovery therefore depends on which layers are exchanged, not on the size of the exchanged slice. The failure of attribution ranking is also not an allocation artefact, because it persists when allocation is fixed.
Specificity and the last six layers. Late layers are effective for Pan+, as our depth sweep showed (Table 20), and the last six layers match specificity there. Specificity reaches this effect from attribution alone, without a depth prior, and with a largely different set, since only layers 37 and 39 are shared. Across directions, specificity also finds effective sets away from the last layers. Its tilt+, tilt− and zoom− sets each share only one layer with layers 34 to 39 and reach median sufficiency ρ of 0.88, 0.98 and 1.00 (Table 1). Specificity therefore does not reduce to a late-layer rule.
Patching sweep. The leave-one-out patching sweep, at 1,680 forward passes, selects a set with lower effect than specificity (0.43 and 0.52).
What is learned. Beyond where the effect is located, controlled exchanges show that it is carried mainly by attention values under the recipient's routing, that it persists only when later sites receive the matching state, and that patched motion follows its own donor's time course (median correlation 0.955 versus 0.020 for other sources).
In the revised paper, we will add these same-grid controls to the main results.