- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper studies hallucination detection through sparse autoencoder (SAE) representations. The proposed framework, HalluSAE, first identifies informative layers using geometric potential energy, then selects sparse features through contrastive logit attribution, and finally trains a linear probe for detection. Experiments on three model families show improved AUROC over the evaluated baselines on curated HaluEval and TriviaQA datasets.
Critical strengths
-
The contrastive attribution strategy is well motivated: examining contributions to incorrect versus correct outputs provides a useful basis for selecting hallucination-related features. The comparisons with unidirectional attribution support this design.
-
The evaluation covers three model–SAE pairs with different SAE architectures. This strengthens the empirical evidence that the approach is applicable beyond a single model.
-
The three-stage framework is easy to follow at a high level. The component ablations and feature-level examples also help explain how the selected representations contribute to detection.
Critical weaknesses
- Interpretation of geometric potential energy. Appendix A defines GPE as the squared distance from a factual centroid, but the interpretation of this statistic as potential energy remains insufficiently justified. In particular, cross-layer distance growth could reflect changes in activation scale or variance rather than a phase transition. Could the authors clarify this distinction and explain how the layer-wise analysis relates to the trajectory in Figure 1?
- Scope of generalization. Although the cross-model evaluation is valuable, both datasets contain predominantly short answers, with an emphasis on numerical and entity errors. Evaluation on longer responses or a different hallucination setting would better support the broader applicability claimed by the paper.
- Causal evidence. The exploratory interventions are useful, but a few examples do not establish that the selected features systematically drive hallucination. Aggregate intervention results with activation-matched controls would strengthen this claim. Otherwise, the paper should distinguish predictive attribution from causal evidence more carefully.
- Limited methodological novelty. Contrastive signals and internal representation hallucination detection have both been explored extensively. HalluSAE appears primarily to combine these established ideas through SAE feature selection and linear probing, while the substantive advance beyond this combination remains unclear.
- Baseline coverage omits similar baselines TSV and ICR prob making the comparison relatively unfair.
Minor issues:
- The introduction’s layout is somewhat cluttered, particularly the text wrapping around Figure 1 Adjusting the figure placement and paragraph spacing would improve readability.
- At the bottom of the first page, “SAPALMA” should be corrected to “SAPLMA.”
Questions for authors
Kindly refer to the questions mentioned in the weaknesses section above.
5 comments
5. Why are TSV and ICR Probe omitted?
- Thank you for identifying these closely related methods. We add TSV and ICR Probe using the same model generations, HaluEval training split, TriviaQA OOD set, and evaluation metrics.† For TSV, we disclose the additional unlabeled-data setting separately; for ICR Probe, we use the official cross-layer configuration and retune only on the training/validation data.
| Method† | Gemma ID/OOD | Llama ID/OOD | Qwen ID/OOD |
|---|---|---|---|
| TSV | 0.855 / 0.762 | 0.824 / 0.748 | 0.846 / 0.751 |
| ICR Probe | 0.873 / 0.781 | 0.851 / 0.772 | 0.869 / 0.778 |
| HalluSAE | 0.929 / 0.804 | 0.887 / 0.801 | 0.908 / 0.793 |
6. Minor presentation issues
- We will reposition Figure 1 to remove the crowded text wrapping, correct "SAPALMA" to "SAPLMA" throughout, fix the duplicated Table 5 numbering, and reconcile the main localization equations with the appendix's implemented start-multiplier and endpoint-tolerance rules.
We thank the reviewers again. The revised paper will make the depth-wise dynamical-system definition explicit, separate predictive attribution from intervention-based causality, and provide matched controls that isolate the contributions of GPE localization, SAE features, and C-DLA.
Thank you for adding these baselines. The reported ICR Probe results appear notably higher than those in the original paper. Could you clarify whether differences in the data splits, training setup, or evaluation protocol might account for this gap?
1. How should GPE and Figure 1 be interpreted, and could the pattern be caused by activation scale?
2. Do the short-answer datasets support broad generalization claims?
3. Do the interventions establish that the selected features systematically drive hallucination?
We agree that two examples alone are illustrative. We therefore add an aggregate bidirectional intervention study using the top-5 positive C-DLA features, activation-matched random features from the same layers, and negative-C-DLA features as controls.
Ablating positive C-DLA features reduces hallucination and increases the correct-vs-incorrect logit margin, whereas amplifying them produces the reverse effect. Activation-matched random interventions have substantially smaller effects.
4. What is methodologically new beyond combining established ideas?
We agree that SAEs, probing, and contrastive attribution are individually established. Our contribution is the specific coarse-to-fine connection between them:
The matched ablation makes the division of labor precise: C-DLA supplies the dominant accuracy gain, while GPE localization supplies an additional matched-budget gain and reduces measured full-layer extraction cost by 68%.† We will revise the novelty statement so that it does not imply that the individual components are new.