- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper revisits feature-wise max pooling for hallucination detection using LLM hidden states. The authors analyze semantic-probability weighting in HaMI and provide theoretical results on margin improvement and pooling behavior. Experiments across five benchmarks and three LLMs show that max pooling outperforms mean pooling and achieves competitive detection AUC without additional sampling.
My main concern is that the theoretical analysis does not fully explain the observed detection improvements. The margin results rely on restricted conditions and do not establish better AUC after training. Moreover, the strong LongFact results rely on automatically generated factuality labels whose reliability has not been independently verified on that benchmark.
Critical strengths
- Strong empirical results with useful architectural ablations.
- Competitive detection performance without additional sampling.
Critical weaknesses
- Limited theoretical explanation for the empirical advantage of max pooling. Theorem 2 establishes a general property of logistic-loss optimization rather than a specific advantage of max pooling. Theorem 3 shows that max pooling avoids gradient dilution under a restrictive sparsity assumption, but the paper does not directly verify whether this assumption holds for the learned token features. More importantly, larger gradients do not necessarily lead to better hallucination detection AUC, and the experiments do not establish that the proposed mechanism accounts for the observed improvements. Given that max pooling is an established architecture, this limits the theoretical contribution of the paper.
- Insufficient validation of the LongFact evaluation labels. The strongest experimental results are obtained on LongFact, where factuality labels are assigned by GPT-5 mini without reference answers or external verification. Although the authors report 95.3% agreement with human annotations on TruthfulQA, this does not establish label reliability for long responses containing multiple factual claims. Furthermore, the 1024-token experiments contain only 69 and 35 hallucinated training examples for the 12B and 70B models, respectively. Without independent validation of the LongFact labels, the reported AUC improvements remain difficult to assess as evidence of genuine factual error detection.
Questions for authors
Is there empirical evidence that the assumed sparse-feature structure holds in the trained detectors and explains the AUC improvement?
Do the LongFact AUC improvements persist when evaluated against independently verified factuality labels?
Could the authors include comparisons with recent single-generation hallucination detectors, such as HaluNet, Simple Factuality Probes, HIDE and Seek, and SinkProbe? Evaluating detection AUC and inference cost under matched settings would help establish the competitiveness of the proposed method.
4 comments
Response to Reviewer P81f (Part2)
W2 & Q2
"Insufficient validation of the LongFact evaluation labels ... the 1024-token experiments contain only 69 and 35 hallucinated training examples."
"Do the LongFact AUC improvements persist when evaluated against independently verified factuality labels?"
Thank you for raising these concerns, which we share. Our original long-form study had two weaknesses that could not be fixed within its design: its labels came from a single LLM judge working from its own knowledge, and its longest cells had only a few dozen hallucinated training answers. We therefore replaced the study with LongFact++ (Obeso et al., 2026), whose labels are independently verified and which is large enough for class-balanced training.
Labels (Section 6.3, Appendix D.2).
- Every entity in an answer is checked against web search by the dataset's annotation pipeline, not by our judge.
- We label an answer hallucinated when at least 10% of its checked entities are not supported.
- The dataset authors validated the pipeline (Obeso et al., 2026). Human annotators who checked a random sample of entity spans with web search agreed with it in 84% of cases (n = 50). On passages with injected factual errors, it detected 80.6% of the errors (729/904), with a 15.8% false-positive rate on unchanged entities.
Training data (Appendix D.2, Table 13). For each of LLaMA-3.1-8B-Instruct and LLaMA-3.3-70B-Instruct (median answer about 740 tokens), we use 1,000 class-balanced training answers (500 hallucinated) and an unweighted loss. Validation has 1,000 answers and test 1,500 at the natural class rates. All methods were retrained under the same protocol.
Answer to Q2: yes, the improvements persist on these labels (Section 6.3, Table 3).
| Method | 8B AUC | 8B AUPRC | 70B AUC | 70B AUPRC |
|---|---|---|---|---|
| SEP | 0.493 | 0.665 | 0.534 | 0.469 |
| LLMs Know | 0.623 | 0.773 | 0.633 | 0.558 |
| SAPLMA | 0.647 | 0.774 | 0.637 | 0.581 |
| HaMI | 0.711 | 0.827 | 0.665 | 0.602 |
| HaMI (SP) | 0.711 | 0.828 | 0.663 | 0.598 |
| Mean | 0.759 | 0.864 | 0.739 | 0.689 |
| Max | 0.783 | 0.881 | 0.739 | 0.691 |
8B: LLaMA-3.1-8B-Instruct; 70B: LLaMA-3.3-70B-Instruct. Mean of five runs. Chance AUPRC equals the hallucination rate (0.674 for 8B, 0.449 for 70B).
- The four MIL detectors outperform the three fixed-position probes on both models.
- Among the methods of Table 3, max pooling is best on LLaMA-3.1-8B-Instruct (AUC 0.783).
- On LLaMA-3.3-70B-Instruct, max pooling ties mean pooling (0.739). It exceeds HaMI by 7.4 points and the fixed-position probes by 10.2–20.5 points.
- Answer length alone reaches only 0.59 and 0.52 AUC (Section 6.3).
Response to Reviewer P81f (Part3)
Q3
"Could the authors include comparisons with recent single-generation hallucination detectors, such as HaluNet, Simple Factuality Probes, HIDE and Seek, and SinkProbe? Evaluating detection AUC and inference cost under matched settings ..."
Thank you for the suggestion. We added all four detectors: HaluNet (Tong et al., 2026), Simple Factuality Probes (SFP; Han et al., 2025), HIDE (Chatterjee et al., 2025) and SinkProbe (Binkowski et al., 2026) (Appendix E.4).
Setting.
- Data and layer search. Same data, splits and labels as max pooling, and the same layer search: every stored layer, validation-selected, five runs.
- SinkProbe and HIDE. SinkProbe uses all layers and HIDE the middle layer, as in their original definitions, with their hyperparameters selected on validation.
- Inference cost. Measured end to end, from the prompt to the score including generation, on the same generated answers (Appendix E.1).
| Method | QA AUC 8B / 12B / 70B (mean of 4 benchmarks) | LongFact++ AUC 8B / 70B | Time overhead vs. generation (median per response, QA) |
|---|---|---|---|
| Max pooling | 0.866 / 0.897 / 0.900 | 0.783 / 0.739 | +0.5% |
| HaluNet | 0.852 / 0.887 / 0.886 | 0.770 / 0.738 | +0.9% |
| SFP | 0.818 / 0.862 / 0.885 | 0.635 / 0.632 | +0.4% |
| SinkProbe | 0.818 / 0.864 / 0.878 | 0.765 / 0.734 | +30.8% |
| HIDE | 0.560 / 0.538 / 0.540 | 0.500 / 0.534 | +8.1% |
- AUC. Among these four recent detectors, max pooling has the highest mean AUC on every generator and on both LongFact++ models (Tables 27 and 28). Per setting, it matches or exceeds every recent detector in 10 of the 12 QA settings; the exceptions are two settings where HaluNet is 0.6 and 1.2 points higher (Table 27).
- Cost. Max pooling, SFP and HaluNet add under 1%. HIDE needs keyword selection with a sentence encoder (+8.1%). SinkProbe must generate with eager attention to obtain attention maps (+30.8% on QA, +15–28% on LongFact++).
- Max pooling on long answers. It needs no stored states: the work after generation stays below 1 ms, while HaluNet and HIDE need up to 1.3 s and 0.9 s on LLaMA-3.3-70B-Instruct (Table 30).
References
- Binkowski, J., Adamczewski, K. & Kajdanowicz, T. (2026). Attention sinks as internal signals for hallucination detection in large language models. ICML.
- Chatterjee, A., Goel, Y. & Chakraborty, T. (2025). HIDE and Seek: Detecting hallucinations in language models via decoupled representations. arXiv:2506.17748.
- Han, J., Band, N., Razzak, M., Kossen, J., Rudner, T. G. J. & Gal, Y. (2025). Simple factuality probes detect hallucinations in long-form natural language generation. Findings of EMNLP.
- Obeso, O., Arditi, A., Ferrando, J., Freeman, J., Holmes, C. & Nanda, N. (2026). Real-time detection of hallucinated entities in long-form generation. arXiv:2509.03531.
- Tong, C., Zhang, Q., Jiang, Z., Jiang, L. & Liu, Y. (2026). HaluNet: Learning hallucination risk from internal signals in LLM question answering. arXiv:2512.24562v2.
Following Up on Our Response to Your Review
Dear Reviewer P81f,
Thank you again for your careful review. In case it is helpful, here is a brief summary of our response.
- W1 / Q1 (theory): Theorem 3 is restated without the sparsity assumption and verified on the trained detectors. We present the gradient-to-AUC link as empirical support, not proof of causation.
- W2 / Q2 (long-form labels): We replaced the study with LongFact++ (retrieval-verified labels). Max pooling is best on LLaMA-3.1-8B and ties mean pooling on LLaMA-3.3-70B.
- Q3 (recent detectors): We added HaluNet, SFP, HIDE, and SinkProbe. Max pooling matches or exceeds them in 10 of 12 QA settings.
We would be grateful to hear whether these address your concerns, or if anything needs further clarification.
Best regards,
The authors
Response to Reviewer P81f (Part1)
We thank the reviewer for the careful review and the useful suggestions. The main changes relevant to this review are:
We address each point below.
W1.1
Thank you for this point; the result is indeed not specific to max pooling. It holds for any differentiable scorer trained with the logistic loss. Its role in the paper is only to show that pooling enters the first-order margin gain through the squared gradient norm ∥∇θzB∥2, the quantity that Theorem 3 compares between max and mean pooling. In the revision we state it as a supporting lemma (Lemma 2) rather than a theorem. The pooling-specific statement is Theorem 3.
W1.2 & Q1
Thank you for these questions. The sparsity assumption is removed and the result is verified on the trained detectors.
On AUC, we agree with the reviewer. A larger gradient does not by itself imply a higher AUC, and Theorem 3 does not claim it; AUC also depends on the test distribution. Lemma 2 links the gradient to the margin, since a bag's first-order margin gain is proportional to its squared gradient norm. We therefore checked each link from gradient to margin to AUC on max- and mean-pooling detectors for the four QA benchmarks.
Every link points in the same direction. We present this as empirical support for the mechanism, not as a proof that the gradient difference causes the AUC difference. The revision states this scope explicitly (Section 5.2).