- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 3(+1)
Read review
Paper summary
Summary This paper introduce Matrix-weighted Attention Graph Energies (MAGE), a framework for extracting attention metrics that can be used as model fingerprints. It represents self-attention as a matrix-weighted graph incorporating attention probabilities and value–output projections. It combines token representations and attention probabilities obtained from a model’s forward pass with its value–output projections, resulting in a single matrix-weighted attention graph captures both the aggregation weights between tokens and the transformations of the transmitted vector information. For model family identification, authors use reference bank comprising 23 models: four Llama, five Qwen, four Gemma, four SmolLM2, four Pythia, and two DeepSeek models. Comparison is done by expressing each prompt’s metric profiles in terms of relative depth in [0, 1] and linearly interpolate them onto 64 uniformly spaced grid points, followed by preprocessing and Dynamic Time Warp matching. Results for within-model profile consistency across prompts shown the consistent layerwise profile shapes recurrence within each model across different prompts. For between model and within family profile structure, authors compute pairwise DTW costs between these model profiles and compare each with the six family representatives used by the classifier, showing that all 23 models had the lowest cost for their own family.For unsee mode family classification, authors classified 30 unseen models while keeping the 23 reference models and six family representative profiles, with results supporting the utility of the three metrics for family identification beyond the reference bank. By applying LoRA-based supervised fine-tuning to eight source models from the Llama and Qwen families using 4,000 instruction–response examples from No Robots and two seeds results , authors demonstrate robustness of family identification under the evaluated LoRA-SFT conditions. Authors analysis points to that cross-prompt consistency of the three metrics and the similarity of profiles within families suggest that MAGE may capture recurring internal model characteristics despite input variation.
Contributions
- paper define three metrics inspired by a matrix-weighted graph representation of self-attention and the Laplacian quadratic energy form.
- paper analyze within-model profile consistency across prompts and profile relationships across models and families
- paper assess the identifying information and scope of the proposed fingerprint through model family classification andbase-lineage attribution for distilled models
Validations
- For within-model cross-prompt consistency, authors evaluate 23 reference models across 6 families (Llama, Qwen, Gemma, SmolLM2, Pythia, DeepSeek) on 1,000 diverse prompts (499,500 prompt pairs per metric).
- For within-family clustering & unseen family classification, authors showed lower median DTW cost within families ( over 35 pairs) than between families ( over 218 pairs) on the 23 reference models (23/23 self-classification).
- For Post-training robustness (LoRA-SFT), authors fine-tuned 8 Llama and Qwen models with LoRA (targeting attention and MLP projections including ) on 4,000 No Robots examples across 2 seeds and 2 epochs (epochs 1 and 3). Achieved 32/32 (100%) accuracy, including when excluding the source checkpoint from the reference bank.
- For Base-lineage attribution on distilled models, authors evaluated 4 DeepSeek-R1-Distill models against their documented Qwen or Llama base families. Correctly matched 3/4 models (Qwen-1.5B, Qwen-7B, Llama-8B), while misclassifying DeepSeek-R1-Distill-Qwen-14B as DeepSeek
Critical strengths
- The paper introduces a clean, practical way to fingerprint LLMs by incorporating the learned value–output projection matrices () directly into attention-graph analysis, rather than relying only on scalar attention probabilities (). By defining three layerwise energy metrics () that handle causal masking and asymmetric weights, MAGE compresses complex forward-pass activations into a compact three-dimensional trajectory across layers. This is consequential because it makes it possible to directly compare models with different hidden dimensions and layer counts using relative-depth interpolation and Dynamic Time Warping (DTW), while capturing identifying signal from the projection weights that improves family classification over scalar attention graphs alone (13/15 vs. 10/15).
- Equally important is the strong empirical evidence that these layerwise profiles are both prompt-invariant and resilient to post-training. Even though attention weights change with every input, testing across 23 reference models and 1,000 diverse prompts shows that a model's layerwise energy curve stays nearly identical regardless of the text fed into it, with median prompt-pair Pearson correlations exceeding 0.949 across all six family representatives in Table 1. For a fingerprint to be useful in practice, it cannot depend on a fragile set of probing queries and must survive downstream adaptation. Showing that forward-pass activations reliably collapse onto a stable, family-specific curve—capable of classifying 25/30 unseen models, holding up across 32/32 LoRA fine-tuned checkpoints, and recovering the base Qwen or Llama lineage for 3 out of 4 distilled DeepSeek models—provides a compelling empirical foundation for white-box model auditing.
Critical weaknesses
- The main weakness of the paper is that it never convincingly shows we actually need dynamic forward-pass attention graphs rather than just simple static weight norms across layers. Because the value-output weights () and normalization layers (LayerNorm or RMSNorm) apply fixed scaling that spikes sharply in the final layers—for example, stays near zero early on before jumping to ~75 in Figure 2b and in Figure 1c—this fixed layer-to-layer jump alone will naturally force any two prompts to have near-perfect Pearson correlations (). While the authors compare MAGE against a scalar attention baseline without , they only test it on 15 of the 30 unseen models using 100 prompts (where 13/15 vs. 10/15 is too small a gap to be convincing), and they completely skip a basic weight-only baseline that doesn't use forward passes. Since MAGE already requires white-box access to the weights and runs up to 1,000 forward passes per model, the core premise of building input-dependent attention graphs is hard to justify if simple weight norms do the same job.
- As defined in paper, distillation evaluation suffers from both an unchecked reference-bank overlap and a misleading motivation. In the LoRA test, the authors carefully verify that removing the exact source model from the 23-model reference bank still gives 32/32 accuracy, yet for the four DeepSeek distilled models (where 3 of 4 match their Qwen or Llama base families), they never mention if the exact base checkpoints (like Qwen2.5-1.5B in Table 1) were removed from the bank. Without that check, it is impossible to tell whether MAGE is picking up a broader family signature or just matching a fine-tuned model back to its exact starting checkpoint. Furthermore, the Introduction motivates the paper with real-world API distillation disputes (like OpenAI/DeepSeek and Anthropic/Moonshot AI), where the goal is catching the teacher model that generated the training data. However, predicting Qwen instead of DeepSeek for DeepSeek-R1-Distill-Qwen shows that MAGE only tracks the student's starting weights and is blind to the teacher—contradicting the opening pitch and leaving the lineage claim resting on just four models.
Questions for authors
- Since MAGE assumes white-box weight access and runs up to 1,000 forward passes per model, how does a purely static, weight-only baseline—such as matching layerwise profiles of or LayerNorm/RMSNorm scale parameters () without running any forward passes or computing attention probabilities —perform on the 30 unseen models?
- In Section 5.4, why was the scalar-weighted Dirichlet energy baseline (setting ) evaluated on only 15 of the 30 unseen models using 100 prompts (yielding 13/15 vs. 10/15) rather than the full 30-model test set? What is the scalar baseline's accuracy across all 30 unseen models and the 4 distilled models?
- In the LoRA-SFT experiment (Section 5.6), authors explicitly verify that excluding the source checkpoint from the 23-model reference bank and recomputing family centroids preserves 32/32 accuracy. For the four DeepSeek distilled models (Table 2), are the exact base checkpoints of the three correctly attributed models (Qwen2.5-1.5B, Qwen2.5-7B, and Llama-3.1-8B) present in the 23-model reference bank while Qwen2.5-14B is absent?
- What are the classification results and DTW costs for the four distilled models when their exact base checkpoints are excluded from the reference bank?
- Because taking a pointwise arithmetic mean across interpolated curves can smear shifted peaks (conflicting with the motivation for DTW in Section 4.2), how does your classifier compare against direct 1-Nearest Neighbor (1-NN) DTW over the 23 reference models, or against simple Euclidean () distance (DTW window = 0) on the 64-point interpolated grid?
- Section 5.5 refers to the 30 unseen models as a "development evaluation". Were pipeline hyperparameters—such as the 64-point interpolation grid, the DTW warping window of 3, and the metric combination ()—selected by maximizing accuracy on these 30 unseen models, or tuned strictly on the 23 reference models?
Suggested revisions:
- The opening paragraph motivates model provenance using disputes over unauthorized black-box API distillation (citing OpenAI/DeepSeek and Anthropic/Moonshot AI), where an auditor wants to identify the teacher model that generated synthetic training data. However, classifying DeepSeek-R1-Distill-Qwen as Qwen rather than DeepSeek shows that MAGE recovers the student's starting base weights and is largely invariant to the distillation teacher. We strongly suggest revising the Introduction so it does not conflate black-box teacher data attribution with white-box base-checkpoint lineage.
- extending the controlled experiments proposed in Section 5.7 to include full-parameter fine-tuning, continual pre-training, and controlled pre-training runs (varying training corpora under a fixed architecture and normalization scheme) would help isolate how much of the MAGE signature stems from learned data representations versus architectural hyperparameters.
8 comments
Response to Reviewer hY1X (Part II)
Responses to Questions
Q1: Our study starts from interpreting token interactions in self-attention as a directed matrix-weighted graph. Our main finding is that energy metrics defined on this representation exhibit recurring layerwise shapes across diverse prompts and similarities among models within the same family. Model-family classification is an application used to evaluate the identifying information and utility of these characteristics. The name MAGE (Matrix-weighted Attention Graph Energies) reflects this perspective.
The edge weight represents both the aggregation weights between tokens, given by attention probabilities, and the vector transformation through the value–output projection. We therefore use to include both components; alone does not represent the input-dependent aggregation weights between tokens.
We also agree that it is useful to test how much weight-only statistics explain profile shapes and family-identification results. We conducted OV-, WV-, and WO-only comparisons using the same 23-model reference bank, 30 unseen models, and comparison procedure. The proposed OV-norm feature averages the headwise Frobenius norms within each layer:
We applied the existing 64-point interpolation and z-score normalization to each profile, averaged model profiles within each family, and compared them using window-3 DTW.
| Method | Input | Correct |
|---|---|---|
| OV norm | Projection weights | 25/30 |
| OV squared norm (secondary control) | Projection weights | 23/30 |
| WV norm | Projection weights | 22/30 |
| WO norm | Projection weights | 20/30 |
| MAGE-100 | 100 prompts | 24/30 |
| MAGE-1000 | 1,000 prompts | 25/30 |
OV norm and MAGE-1000 each correctly classified 25 of the 30 models. However, MAGE correctly classified all five models that OV norm misclassified, and OV norm correctly classified all five that MAGE misclassified. Thus, the same overall accuracy was accompanied by different model-level family predictions. We added the feature definitions, comparison procedure, and predictions for all models to Appendix I.5, Tables 30–31.
Q2: We designed the scalar comparison as a supporting controlled experiment to examine the contribution of the value–output transformation in the matrix-weighted graph. Profiles with this transformation replaced by the identity matrix require separate extraction, so extending the evaluation to all 30 queries entails additional computation time. Given this experimental purpose and the available time and computing resources, we retained the existing 23-model reference bank and conducted the comparison on 15 unseen queries using 100 prompts.
Both methods used the same selected models, prompts, preprocessing, and DTW procedure. Under these conditions, the scalar baseline obtained 10/15 and MAGE obtained 13/15. Appendices I.1–I.2, Tables 26–27, report the model-level results and experimental details. Scalar accuracies for all 30 queries and the four distilled models have not yet been computed.
Q3: The exact starting checkpoints, checked against the official model card, are:
| Distilled query | Documented exact source | In the original 23-model bank? |
|---|---|---|
| DeepSeek-R1-Distill-Qwen-1.5B | Qwen2.5-Math-1.5B | Yes |
| DeepSeek-R1-Distill-Qwen-7B | Qwen2.5-Math-7B | No |
| DeepSeek-R1-Distill-Llama-8B | Llama-3.1-8B | No |
| DeepSeek-R1-Distill-Qwen-14B | Qwen2.5-14B | No |
Only the Qwen-1.5B query's exact source is present. Qwen-7B and Llama-8B were correctly classified as Qwen and Llama even though their starting checkpoints were absent from the bank. The absence of the source therefore applies to these two correctly classified models as well as Qwen-14B. Appendix B.1, Table 5, lists the references; Appendix G.3, Table 17, lists the documented starting checkpoints.
Response to Reviewer hY1X (Part III)
Q4: For each query, we removed its own exact source wherever present and recomputed the affected family representative, keeping the other references, preprocessing, and DTW settings fixed. The Qwen-1.5B query consequently uses 22 references; the other three use the original 23, which already exclude their respective sources.
The full family DTW costs under this query-specific exclusion are below. The common “DeepSeek-R1-Distill-” prefix is omitted. The smallest cost determines the prediction.
| Query | Bank size | Prediction | Llama | Qwen | Gemma | SmolLM2 | Pythia | DeepSeek |
|---|---|---|---|---|---|---|---|---|
| Qwen-1.5B | 22 | Qwen | 1.037912 | 0.150101 | 2.129558 | 0.817432 | 1.922925 | 1.002349 |
| Qwen-7B | 23 | Qwen | 1.613485 | 0.855026 | 2.129357 | 1.684006 | 1.970548 | 0.924934 |
| Llama-8B | 23 | Llama | 0.163807 | 1.007428 | 1.932940 | 0.557134 | 1.192490 | 0.431525 |
| Qwen-14B | 23 | DeepSeek | 0.980476 | 0.790878 | 2.113479 | 0.865643 | 1.830309 | 0.477076 |
For Qwen-1.5B, the Qwen cost increases from 0.104649 to 0.150101 while the prediction is retained. The other queries' banks and costs are unchanged. All four predictions and the 3/4 correct count therefore remain the same under query-specific exact-source exclusion.
Q5: Using the existing 1,000-prompt profiles of the same 23-model bank and 30 unseen models on a 64-point grid, we obtained the following results.
| Representatives | Distance | Correct |
|---|---|---|
| Six family-average profiles | DTW, | 25/30 |
| 23 individual references, 1-NN | DTW, | 22/30 |
| Six family-average profiles | Squared Euclidean, | 19/30 |
The 1-NN result comes from reanalyzing the existing profiles; the family-average DTW and Euclidean results are reported in Appendix H.3, Table 23.
Q6: The 23-model bank, , , and default combination were specified before observing the classification results for the 30 unseen models. We constructed the classifier using reference-bank profiles and applied it to the 30 external checkpoints with those choices fixed.
We subsequently reused the collected profiles to analyze sensitivity to the grid, window, metric combination, and number of prompts, as reported in Appendix H. We retained the original configuration after these analyses; the main-text result of 25/30 uses that configuration.
The terms “development evaluation” and “development panel” could unintentionally suggest that we selected the settings based on the 30-model results. We therefore revised them in the Abstract, Section 5.5, and Appendix H to distinguish evaluation with a fixed configuration from the subsequent sensitivity analyses.
Response to Reviewer hY1X (Part I)
Thank you for your careful review.
We revised the manuscript to make the progression clearer. MAGE represents token interactions in self-attention as a directed matrix-weighted graph. We observe recurring layerwise energy profiles across diverse prompts and similarities within model families, then evaluate their identifying information through model-family classification.
The tail-layer, query-specific exact-source, and direct-reference analyses are reported directly below.
Responses to Weaknesses
W1: We performed the requested weight-only comparison and explain its results and relation to our research perspective in Q1. Here we address how much the final-layer peaks affect the high Pearson correlations.
Using the existing 1,000-prompt profiles of all 23 reference models, we removed either the last layer or the last ⌈L/10⌉ layers and recomputed Pearson correlations. We used the original layer values before interpolation and compared all 499,500 prompt pairs for each model and metric. The table reports the median across the 23 models of their prompt-pair Pearson medians.
For Llama-3.2-3B in Figure 2, the Dl median is 0.9968 with all layers, 0.9967 after removing the last layer, and 0.9974 after removing the last three. High consistency therefore persists in the remaining profile in this example. Some models show decreases: Pythia-410M's Sl falls from 0.9429 to 0.8839 after removing the final 10%. Overall correlations remain high while exhibiting these model-specific changes.
We used 1,000 prompts to examine profile consistency across diverse inputs. Appendix H.1, Table 21, reports classification with fewer prompts. Across 20 trials using the same prompt subset for the bank and queries, 10 prompts yielded an average of 24.00/30 correct predictions, and 100 prompts yielded 24.65/30.
W2: We checked which starting checkpoints are present in the bank and evaluated each query with its own exact source excluded. All four predictions and the 3/4 accuracy were retained. Q3–Q4 provide the checkpoint list, exclusion conditions, and all family DTW costs.
The examples in the Introduction motivate examining how models were developed and where they originated. The revised text connects this motivation to studying internal self-attention characteristics when weights and attention are accessible. The DeepSeek experiment asks whether the original base-family characteristics remain in models initialized from Qwen/Llama checkpoints and trained on data generated by DeepSeek-R1. Section 5.6 and Appendix G.3 identify the documented starting checkpoints as the basis for the target labels and report base-lineage attribution for these four models.
The suggested full-parameter fine-tuning, continual pre-training, and controlled pre-training experiments would help examine how training affects the energy profiles. However, varying training data while fixing architecture and normalization requires models with matched initialization, training procedures, and other conditions. If suitable public checkpoints cannot be obtained, the comparison models need to be trained specifically for this purpose. Constructing and training such a controlled model set was beyond what we could complete during this revision. Section 5.7 and the Conclusion describe the corresponding research direction: separating the effects of training and model design and tracking intermediate training checkpoints.