- Rating
- 1
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper investigates whether the workspace (J-space), previously identified as carrying verbalizable content, also mediates control decisions in language models, including stopping generation, ending reasoning, tool calling, and refusal. The authors construct steering directions associated with control behaviors by computing differences in mean residual-stream activations. They then test whether the effects of these directions on model outputs persist when J-space activations are clamped to their values from the unperturbed run. The authors also use a content-related steering direction as a positive control. Experiments on Qwen3-4B show that clamping J-space substantially reduces the effect of content-related steering (65% mediation), but has almost no effect on steering the stop-generation token. The authors further observe similar patterns across different control tokens and model architectures, and investigate how this phenomenon changes across different stages of post-training. Based on these findings, the authors argue that control-related signals can influence model outputs without passing through the identified content workspace, and discuss the potential implications for chain-of-thought monitoring.
Critical strengths
-
The research question is clear and relevant. The paper raises a well-defined and worthwhile question: whether the J-space previously identified for verbalizable content also participates in the model's control decisions. Understanding how language models make decisions such as stopping generation, calling tools, and refusing requests is important for both mechanistic interpretability and model transparency research. I believe this is a worthwhile research question to explore.
-
The experimental evaluation covers a relatively broad range of settings. The authors conduct experiments across multiple model architectures and different types of control decisions. In addition, the paper includes a content-related positive control and performs a series of robustness checks, such as varying the steering strength and using different geometric metrics.
These experiments are valuable because they help determine whether the observed phenomenon is specific to a particular model or experimental setting. However, these experiments alone are not sufficient to fully establish the mechanistic explanation proposed by the authors.
Critical weaknesses
- The main experiments do not sufficiently support the paper's central mechanistic conclusion. The paper claims that control decisions are computed outside J-space. However, the main experiments actually test whether the effects produced by artificially constructed steering directions persist after J-space is clamped. For example, the stop direction is constructed by comparing activations at the stopping position in a reasoning trace with those at the midpoint of the thinking block. It is currently unclear whether this direction captures the actual stopping decision or merely reflects other differences between these two positions. More importantly, demonstrating that an artificially injected direction can influence the stop-token logit without passing through J-space does not establish that the model's stopping decision during natural operation is also computed outside J-space. Although the appendix provides some additional ablation evidence, this issue remains insufficiently addressed, particularly for the other types of control decisions.
This is a fundamental concern because the paper's central conclusion relies on the assumption that the results of the steering experiments reflect the model's actual underlying decision-making mechanism.
- The effectiveness of the J-space intervention has not been sufficiently validated across different experimental settings. The paper defines the workspace using sparse gradient pursuit, which depends on the current activation state, but most experiments actually use a fixed linear approximation. Section 4.1 examines the actual decomposition, but it remains unclear whether the core clamp-mediation results can be reproduced using the full decomposition method. Furthermore, the content positive control used to validate the effectiveness of the clamp is typically evaluated on a different task distribution from that used in the control-decision experiments. This matters because, in the tool-calling task, the specially constructed content positive control was not clearly validated (mediation of 0.24, compared with 0.33 for the random-direction control). The positive control also failed validation on OLMo-3-Think. Without more thorough validation of the intervention method in the corresponding experimental settings, it is difficult to determine to what extent the reported bypass results can be generalized.
Questions for authors
-
How do the authors distinguish an artificial J-space bypass produced by steering from the actual computational pathway used for control decisions during the model's natural operation? Specifically, the stop direction is constructed by computing the difference in activations between the stopping position in a reasoning trace and the midpoint of the thinking block. What evidence demonstrates that this direction genuinely captures the computational mechanism underlying the stopping decision, rather than other differences between these two positions, or simply a direction that can directly influence the stop-token readout?
-
How do the authors validate the effectiveness of the identified J-space across different model architectures and control tasks? Section 3 states that each model uses a Jacobian Lens fitted through layer 34. However, some of the evaluated model architectures, such as Llama-3-8B (32 layers) and GPT-OSS-20B (24 layers), have fewer than 34 layers in total. How exactly is the Jacobian Lens constructed for these models? How do the authors ensure that the resulting workspaces are comparable across different models? Furthermore, most experiments use a fixed linear approximation rather than the activation-dependent sparse-gradient-pursuit decomposition. Do the core clamp-mediation results also hold under the full decomposition method? For experimental settings in which the content positive control was not clearly validated or directly failed validation, such as tool-calling and OLMo-3-Think, how should the authors interpret the corresponding bypass results?
Presentation and Reproducibility Issues:
I noticed several issues related to formatting, experimental reporting, and references that need to be addressed.
-
Formatting issue (Page 9): The page number overlaps with the text at the end of the Limitations section.
-
Empty Reproducibility Statement (Page 9): The Reproducibility Statement section is completely empty. Although this section is optional under ICLR guidelines, the authors should either provide relevant reproducibility information or remove the empty heading.
-
Table formatting: The captions for Tables 1–5 are all placed below the tables, which does not comply with the ICLR 2027 formatting requirements. ICLR 2027 explicitly requires table captions to be placed above the tables.
Bibliographic Inaccuracies:
The reference list also contains numerous inaccuracies in author names and paper titles. By comparing the references in the submission with their original sources, I identified the following discrepancies.
| Reference | Information in the submission | Correct information | Source |
|---|---|---|---|
| Darade & Thorat (2026) | Authors: Snehal Darade, Sushant Thorat | Authors: Ishita Darade, Sushrut Thorat | arXiv:2605.22488 |
| Fadnavis et al. (2026) | Third author: Rafael Wyss | Third author: Felix Wyss | arXiv:2605.29126 |
| Sun et al. (2026) | First author: Chuan-En Sun | First author: Chung-En Sun | arXiv:2605.09252 |
| Jiang et al. (2026) | First author: Xu Jiang; Title: A Value Axis in Reasoning Models | First author: Nick Jiang; Title: The Value Axis: Language Models Encode Whether They're on the Right Track | arXiv:2606.17056 |
| Baherwani et al. (2026) | First author: Vasu Baherwani | First author: Vatsal Baherwani | arXiv:2607.22925 |
| Mazaheri & Mazaheri (2026) | Authors: Parisa Mazaheri, Kian Mazaheri | Authors: Parsa Mazaheri, Kasra Mazaheri | arXiv:2605.20530 |
| Wang et al. (2026) | First author: Zhen Wang; Title: New Skills or Sharper Primitives? | First author: Zhilin Wang; Full title: New Skills or Sharper Primitives? A Probabilistic Perspective on the Emergence of Reasoning in RLVR | arXiv:2602.08281 |
| Wu et al. (2026a) | Authors: Jie Wu, Yue Xie, Sheng Lin, Song Zhao, Xin Chen | Authors: Jinman Wu, Yi Xie, Shen Lin, Shiqian Zhao, Xiaofeng Chen | arXiv:2603.05773 |
| Wu et al. (2026b) | First author: Yue Wu; Title: Tool Calling is Linearly Readable and Steerable in Large Language Models | First author: Zekun Wu; Title: Tool Calling is Linearly Readable and Steerable in Language Models | arXiv:2605.07990 |
| Galeone et al. (2026) | Title: Perfect Detection, Failed Control | Full title: Perfect Detection, Failed Control: The Geometry of Knowing vs. Steering in Language Models | arXiv:2606.24952 |
| Dutta (2026), Footnote 1 | Author: S. Dutta | Author: Chandram Dutta | LessWrong article |
These errors do not directly invalidate the paper's experimental results. However, such a large number of bibliographic inaccuracies raises concerns about the level of rigor exercised by the authors during the preparation and checking of the manuscript. I recommend that the authors carefully review the entire reference list and correct these errors.