Pith. sign in

REVIEW 5 major objections 6 minor 32 references

STEP Planner: Constructing cross-hierarchical subgoal tree as an embodied long-horizon task planner

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read STEP constructs a hierarchical subgoal tree through two closed-loop LLM models, claiming 40% success on the WAH-NL long-horizon benchmark and 6/10 on long-complex real-robot tasks, up from 6% and 1/10 for the strongest baselines.

desk verdict Read the tables, not the abstract: the 34%/25% figures are percentage-point improvements over LoTa-Bench, not absolute success rates, and the core claim of a real gain from subgoal-tree decomposition with an LLM termination check is plausible but needs a few missing baselines and artifacts. read the letter →

arxiv 2506.21030 v2 pith:F363MA5C submitted 2025-06-26 cs.RO

Tianxing Zhou , Zhirui Wang , Haojia Ao , Guangyan Chen , Boyang Xing , Jingwen Cheng , Yi Yang , Yufeng Yue This is my paper
classification cs.RO
keywords long-horizontaskplanningsubgoaltreelargelanguagemodelsembodiedagentsrobothierarchicaldecompositionleafnodeterminationVirtualHome
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that long-horizon embodied planning becomes reliable when a large language model never has to reason about the whole task at once. Instead, STEP recursively decomposes a natural-language goal into a subgoal tree, and only stops decomposing a branch when a separate termination model judges the current subgoal to be directly executable as a primitive action. The authors argue this shrinks two failure sources: the contextual gap, where long instruction and action histories add irrelevant information that degrades LLM reasoning, and the logical gap, where abstract instructions are too far from concrete actions for one inference step. If the claim is right, then the route to better long-horizon robot planners is not a stronger planner but a structure that feeds each reasoning step only the information it needs. The paper reports head-to-head gains on the VirtualHome WAH-NL benchmark and on a real Franka Panda robot.

What carries the argument

The central object is the subgoal tree: a hierarchy whose root is the natural-language goal, each child node is a subgoal that contributes to its parent's effect, and leaf nodes are primitive actions. Spinning up this tree is done by two closed-loop models: the subgoal decomposition model, which generates the next subgoal from the parent node, the previously generated subgoal at the same level, and the current observation; and the leaf node termination model, which evaluates each subgoal against mappability criteria (can it map to a distinct primitive action?) and consistency criteria (does it match the embodiment's affordances, agree with task congruence, and fit the current environmental constraints?). The termination model's verdict decides whether to execute the action, decompose further, or replan the parent. This architecture confines each LLM call to a single-level decomposition, excluding history from other branches and thereby reducing both the contextual and logical gaps identified in the paper.

What would settle it

Run the WAH-NL benchmark with a version of STEP where leaf-node termination is decided by a deterministic ground-truth check (a subgoal is terminal only if it exactly matches a primitive action known to be executable in the current state) instead of the LLM's mappability/consistency verdict; if this deterministic version does not reproduce or exceed the reported 40% success rate, then the framework's claimed success rides on the LLM termination judgments rather than on the tree structure itself.

Watch

Extended reading notes

Core claim

The central discovery is that a subgoal tree built by recursive decomposition, with a closed-loop leaf-node termination model deciding when to stop decomposing, substantially outperforms direct LLM planning and existing tree-search planning baselines on long-horizon embodied tasks. On the WAH-NL benchmark, STEP reaches 40% overall task success and 62.02% subgoal success, versus 6% and 36.79% for LoTa-Bench, and on real-robot tasks it completes 6/10 long-complex trials versus 1/10 for all baselines. The paper's error analysis attributes its remaining failures mainly to Additional/Missing Steps, which stem from misjudgments by the LLM-based leaf-node termination model. Ablations show that removing the tree structure drops success to 8%, and removing the subgoal condensation drops it to 9%, supporting the claim that both contextual-gap reduction and logical-gap reduction contribute to the gains.

Load-bearing premise

The whole method depends on the LLM correctly deciding when a subgoal is already a primitive action; if it says 'mappable' too early, an invalid action is executed, and if it says 'not mappable' too often, the planner over-decomposes or replans unnecessarily, with no verification against ground truth.

Editorial extensions

If this is right

  • If STEP's claim holds, LLM-based planners can handle significantly longer tasks by replacing full-horizon reasoning with per-level decomposition, without requiring a stronger backbone model.
  • The reported error distribution points to the leaf-node termination model as the main bottleneck, so improving mappability and consistency judgments should directly raise overall success rates.
  • The real-robot results suggest the advantage grows with task complexity, since the largest gap over baselines appears in long-complex tasks involving hidden objects and 5-8 step sequences.
  • Ablation results imply that both the tree structure's context condensing and the subgoal reasoning condensation are individually necessary, not just helpful, for the reported performance.
  • Because the grammar-error rate stays similar to baselines using the same LLM, the gains are attributable to the planning structure rather than to superior language generation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper's claims, the subgoal-tree idea could be combined with sampling or search at each decomposition level, e.g., generating several candidate subgoal sets and selecting among them, which the paper does not explore.
  • A testable extension the paper leaves open is replacing the LLM termination judgment with a learned feasibility classifier or a simulator-based check; if such a replacement lifts success further, it would confirm the termination model as the bottleneck.
  • The contextual-gap argument suggests a measurable prediction: success rate should correlate inversely with the length of the context actually fed to each decomposition call, independent of total task length, which could be tested by varying context condensation.
  • The decomposition-into-subgoal-tree pattern might transfer beyond text-based planning to vision-language-action policies, where each subgoal could condition a different low-level policy or skill instead of a primitive action.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. The paper proposes STEP, an LLM-based embodied long-horizon planner that recursively decomposes a natural-language goal into a subgoal tree. A subgoal decomposition model expands nodes, and a leaf-node termination model uses mappability and consistency criteria to decide whether a node is executable as a primitive action, should be decomposed further, or should trigger replanning. The system is evaluated on VirtualHome WAH-NL (100 annotated tasks) against SayCan, ProgPrompt, and LoTa-Bench, and on a Franka Panda with RoboScript against the same baselines plus RoboScript. The paper reports that STEP reaches 40% success rate on WAH-NL versus 6% for LoTa-Bench, and 6/10 versus 1/10 on long-complex real-robot tasks, with ablations showing that the tree structure contributes most of the gain.

Significance. The central idea is timely: recursively decomposing instructions into a coarse-to-fine subgoal tree is a natural way to reduce both context length and the logical distance between instruction and action. The method has no trained parameters and is evaluated against an external benchmark with objective subgoal-satisfaction criteria, which lowers circularity concerns. If the reported improvements survive a corrected comparison and more careful real-robot statistics, the result would be a useful and simple planning framework. However, the current paper's quantitative claims are obscured by inconsistent presentation, missing comparisons, and a small real-robot sample.

major comments (5)
  1. [Abstract and Section I, Contribution (III)] The abstract and contributions report 'success rates up to 34% (WAH-NL) and 25% (real robot)', while Table I reports STEP at 40% SR and Table III reports 6/10 in the long-complex real-robot category. Section IV-A clarifies that 34 percentage points and 25.23 percentage points are improvements over LoTa-Bench (40-6=34; 62.02-36.79=25.23), and the real-robot aggregate is also a 25 percentage point improvement (87.5% minus 62.5% across the four categories). Because the abstract does not say 'improvement', the headline numbers are internally inconsistent as written. The abstract and contributions should be rewritten to state the base rates and the improvement explicitly.
  2. [Section IV-B, Real Robot] The real-robot evaluation uses two tasks per category with five trials each, giving n=10 per condition and no confidence intervals or significance tests. For the headline long-complex comparison, STEP's 6/10 versus LoTa-Bench's 1/10 yields a Fisher exact p-value of approximately 0.03 one-sided and about 0.06 two-sided; with only two distinct tasks per category this is not a robust statistical basis for the claim of superiority. Please report the raw per-task records, add confidence intervals, and perform a suitable significance test or explicitly weaken the conclusion.
  3. [Section IV-A, Baselines] The baseline set omits Tree-Planner and RAP, the tree-structured planners discussed in Section II-B as the closest related work. Without evaluating against these methods or explaining why they are excluded, the statement that STEP 'outperforms SOTA methods' (Abstract and Section I) is not supported. The authors should add these baselines to the VirtualHome comparison or restrict the claim to the compared set.
  4. [Section III-C, Algorithm 1] Algorithm 1 does not implement the three-way leaf-node decision described in Section III-C. The text states that a non-mappable but consistent subgoal should be further decomposed, while an inconsistent subgoal should trigger replanning; Algorithm 1 instead routes every non-mappable result to line 13 (subgoal = subgoal.parent) and never adds the newly decomposed subgoal to the tree. This makes the algorithm impossible to execute as written and prevents reproduction. Replace Algorithm 1 with a correct procedure and make the tree construction explicit.
  5. [Section III-C, Eqs. (4)-(7)] The formal criteria in Eqs. (4) through (7) are not well-formed. Variables such as ai, T, Tt, and ΦT are used without definitions, and the formulas state injectivity or containment conditions rather than the mappability and consistency predicates they claim to define. Since the termination model is the central control loop, the authors should either specify the actual LLM evaluation protocol (for example, prompt template and output schema) or clearly mark these equations as illustrative rather than formal definitions.
minor comments (6)
  1. [Section III-A] There are typos such as 'permitive action' (should be 'primitive action') and 'enabliing' in Section I; the manuscript would benefit from a careful proofread.
  2. [Section III-B, Eq. (2)-(3)] The notation Φn({bi}|n i=0) is used with braces and subscripts that are not defined consistently; please rewrite the decomposition operator notation so that the recursion is unambiguous.
  3. [Table II] The error-type percentages in Table II do not sum to 100 (for example, the STEP row sums to 37%), and the 15% Grammar Error rate discussed in the text appears nowhere in the table. Please state the denominator and include grammar errors as a column.
  4. [Section IV-A, Dataset] The description 'annotated dataset of 100 entries in the WAH-NL benchmark, which consists of 350 labeled robot datasets' is ambiguous; clarify whether 100 tasks or 350 episodes were evaluated and how the two numbers relate.
  5. [Reproducibility] No code, prompts, LLM version for VirtualHome, or sampling parameters are provided; please include these in the paper or supplement, as the method is otherwise difficult to reproduce.
  6. [References] Reference [8] is a duplicate of reference [7] and does not appear to cite Tree-Planner as intended; please correct the citation list in Section II-B.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: STEP's derivation is self-contained and externally evaluated; reported number inconsistencies are a presentation issue, not a circular step.

full rationale

STEP's claimed derivation—that recursive decomposition plus leaf-node termination yields executable actions—is not circular. The subgoal decomposition model (Eqs. 2–3) is a recursive LLM procedure with no fitted parameters; the leaf termination model (Eqs. 4–7) is an LLM judgment rule, but the reported success rates are not defined by that judgment. On WAH-NL, success is determined by the benchmark's independent subgoal-completion criteria; on the real robot, success is reported as per-task counts over four categories. No parameter is fitted to the test data, and no 'prediction' is statistically forced by an input subset. The paper's self-citations ([4], [10], [12], [19]) appear only in related work and are not load-bearing for STEP's architecture or results. The abstract/contribution figures '34% (WAH-NL) and 25% (real robot)' are likely intended as improvements over LoTa-Bench (40% - 6% = 34 percentage points; average real-robot success 87.5% vs 62.5% = 25 percentage points), so the apparent conflict with Table I (40%) and Table III (6/10) is a wording/presentation inconsistency rather than a circular derivation. The LLM-based leaf termination model's self-judgment is a limitation acknowledged by the authors (Additional/Missing Step errors, Table II), but it does not make the central evaluation circular because the final success metric is external to the planner's internal mappability verdicts.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The central claim rests on assumptions about LLM reliability, subgoal compositionality, and observation sufficiency, but no numeric free parameters are fitted. The invented-entities list is empty because the subgoal tree is a representational structure, not a new physical entity.

assumptions (3)
  • domain assumption The LLM decomposition model reliably produces valid, semantically correct subgoals from parent goals.
    The entire method depends on GPT-4o decomposing goals into coherent subgoals without formal guarantees. Invoked in Sec. III-B.
  • domain assumption Executing all leaf-node actions in the subgoal tree achieves the root goal (compositionality).
    The method assumes that the cumulative effect of all leaves equals the effect of the root. Stated in Sec. III-A without proof.
  • domain assumption Current observation o_t is sufficient to decide both subgoal decomposition and termination.
    The algorithm uses only the current observation and the parent/left-side subgoals; no history or full state is used. Seen in Eq. (2), Eq. (3) and Algorithm 1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of STEP Planner: Constructing cross-hierarchical subgoal tree as an embodied long-horizon task planner." pith.science (2026). https://pith.science/paper/F363MA5C

@misc{pith2026250621030,
  author       = {Pith},
  title        = {Pith review of: STEP Planner: Constructing cross-hierarchical subgoal tree as an embodied long-horizon task planner},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F363MA5C}},
  note         = {Machine review of arXiv:2506.21030}
}
read the original abstract

The ability to perform reliable long-horizon task planning is crucial for deploying robots in real-world environments. However, directly employing Large Language Models (LLMs) as action sequence generators often results in low success rates due to their limited reasoning ability for long-horizon embodied tasks. In the STEP framework, we construct a subgoal tree through a pair of closed-loop models: a subgoal decomposition model and a leaf node termination model. Within this framework, we develop a hierarchical tree structure that spans from coarse to fine resolutions. The subgoal decomposition model leverages a foundation LLM to break down complex goals into manageable subgoals, thereby spanning the subgoal tree. The leaf node termination model provides real-time feedback based on environmental states, determining when to terminate the tree spanning and ensuring each leaf node can be directly converted into a primitive action. Experiments conducted in both the VirtualHome WAH-NL benchmark and on real robots demonstrate that STEP achieves long-horizon embodied task completion with success rates up to 34% (WAH-NL) and 25% (real robot) outperforming SOTA methods.

Figures

Figures reproduced from arXiv: 2506.21030 by the authors.

Figure 1
Figure 1. Illustration of our STEP. Through constructing a hierarchical tree structure that decomposes tasks from coarse to fine granularity, STEP effectively bridges both logical and contextual gaps in the planning process, demonstrating superior performance compared to existing methods in long-horizon and complex tasks. Abstract— The ability to perform reliable long-horizon task planning is crucial for deploying robots in r… view at source ↗
Figure 2
Figure 2. Framework of STEP. a) STEP constructs a subgoal tree through a closed-loop process of subgoal decomposition and leaf-node termination. In the subgoal decomposition model, the LLM breaks down complex tasks into subtasks. Each subgoal is then evaluated against mappability and consistency criteria to determine whether it should be further decomposed, re-planned, or executed directly. b) STEP progressively refines subgo… view at source ↗
Figure 3
Figure 3. Framework of subgoal decompose model. Taking decomposing “Storing tools in drawer” as an example, the subgoal decomposition model can directly infer the next subgoal to be generated from the parent node and the previous step, i.e., grasp tape. capitalizing on the parent task’s inherent encompasses of higher-order mission objectives, result in lower logical gaps. Consequently, the hierarchical decomposition framework… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Framework of leaf node termination model. Mappability criteria and consistency criteria are used to determine whether the current node should execute a primitive action, execute the next decomposition, or replan. The logical structure of the criterion is illustrated in…
Figure 5
Figure 5. Figure 5: Planning results. The experiment was tested on the Virtual Home simulator and real robots, demonstrating the planning performance of STEP in complex environments and long tasks. TABLE II ERROR TYPE ANALYSIS OF DIFFERENT METHODS ON WAH-NL BENCHMARK. Method Goal Satisfac…
Figure 6
Figure 6. Figure 6: Task Length and Subgoal Success Rate Variation under Different Baselines. In the WAH benchmark, we group tasks by length into 10% intervals, from short to long, and calculate the subgoal success rate for each task within the intervals. using a subgoal decomposition mod…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

32 extracted references · 18 canonical work pages

  1. [1]

    Mu Yao, Zhang Qinglong, Hu Mengkang, and et. al. Embodiedgpt: Vision-language pre-training via embodied chain of thought. Advances in Neural Information Processing Systems , 36, 2024

    work page 2024

  2. [2]

    Chen Guanyan, Wang Meiling, Cui Te, and et. al. Vlmimic: Vision language models are visual imitation learner for fine-grained actions. arXiv preprint arXiv:2410.20927 , 2024

    work page Pith review arXiv 2024

  3. [3]

    Wake Naoki, Kanehira Atsushi, Sasabuchi Kazuhiro, and et. al. Gpt-4v (ision) for robotics: Multimodal task planning from human demonstration. IEEE Robotics and Automation Letters , 2024

    work page 2024

  4. [4]

    Fmimic: Foundation models are fine-grained action learners from human videos

    Guangyan Chen, Meiling Wang, Te Cui, Yao Mu, Haoyang Lu, Zicai Peng, Mengxiao Hu, Tianxing Zhou, Mengyin Fu, Yi Yang, and Yufeng Yue. Fmimic: Foundation models are fine-grained action learners from human videos. arXiv preprint, 2025

    work page 2025

  5. [5]

    Sermanet Pierre, Ding Tianli, Zhao Jeffrey, and et. al. Robovqa: Multimodal long-horizon reasoning for robotics. In 2024 IEEE International Conference on Robotics and Automation (ICRA) , pages 645–652. IEEE, 2024

    work page 2024

  6. [6]

    S P Sharan, Zhao Ruihan, topcu ufuk, and et. al. Plan diffuser: Ground- ing llm planners with diffusion models for robotic manipulation. In Bridging the Gap between Cognitive Science and Robot Learning in the Real World: Progresses and New Directions , 2024

    work page 2024

  7. [7]

    Yao Shunyu, Yu Dian, Zhao Jeffrey, and et. al. Tree of thoughts: Deliberate problem solving with large language models. Advances in neural information processing systems , 36:11809–11822, 2023

    work page 2023

  8. [8]

    Yao Shunyu, Yu Dian, Zhao Jeffrey, and et. al. Tree of thoughts: Deliberate problem solving with large language models. Advances in Neural Information Processing Systems , 36, 2024

    work page 2024

Show all 32 references
  1. [9]

    Hao Shibo, Gu Yi, Ma Haodi, and et. al. Reasoning with language model is planning with world model. arXiv preprint arXiv:2305.14992, 2023

    2023 arXiv

  2. [10]

    Unifying latent action and latent state pre-training for policy learning from videos

    Guangyan Chen, Meiling Wang, Te Cui, Luojie Yang, Qi Shao, Lin Zhao, Tianle Zhang, Yihang Li, Yi Yang, and Yufeng Yue. Unifying latent action and latent state pre-training for policy learning from videos. arXiv preprint, 2025

    2025

  3. [11]

    pi 0: A vision-language-action flow model for general robot control

    Kevin Black, Noah Brown, Danny Driess, Adnan Esmail, Michael Equi, Chelsea Finn, Niccolo Fusai, Lachy Groom, Karol Hausman, Brian Ichter, et al. pi 0: A vision-language-action flow model for general robot control. arXiv preprint arXiv:2410.24164 , 2024

    arXiv

  4. [12]

    Graphmimic: Graph-to-graphs generative modeling from videos for policy learning

    Guangyan Chen, Te Cui, Meiling Wang, Chengcai Yang, Mengxiao Hu, Haoyang Lu, Yao Mu, Zicai Peng, Tianxing Zhou, Xinran Jiang, Yi Yang, and Yufeng Yue. Graphmimic: Graph-to-graphs generative modeling from videos for policy learning. In Proceedings of the IEEE/CVF Conference on ...

    2025

  5. [13]

    Vemprala Sai, Bonatti Rogerio, Bucker Arthur, and et. al. Chatgpt for robotics: Design principles and model abilities. IEEE Access , 2024

    2024

  6. [14]

    Yao Shunyu, Zhao Jeffrey, Yu Dian, and et. al. React: Syner- gizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629, 2022

    2022 arXiv

  7. [15]

    Ahn Michael, Brohan Anthony, Brown Noah, and et. al. Do as i can, not as i say: Grounding language in robotic affordances. arXiv preprint arXiv:2204.01691, 2022

    2022 arXiv

  8. [16]

    Shinn Noah, Cassano Federico, Berman Edward, and et. al. Reflexion: Language agents with verbal reinforcement learning. Advances in Neural Information Processing Systems , 36, 2024

    2024

  9. [17]

    Lin Bill, Yuchen, Fu Yicheng, Yang Karina, and et. al. Swiftsage: A generative agent with fast and slow thinking for complex interactive tasks. Advances in Neural Information Processing Systems , 36, 2024

    2024

  10. [18]

    Huang Wenlong, Xia Fei, Xiao Ted, and et. al. Inner monologue: Embodied reasoning through planning with language models. arXiv preprint arXiv:2207.05608, 2022

    2022 arXiv

  11. [19]

    Cui Te, Chen Guangyan, Zhou Tianxing, and et. al. Human demon- strations are generalizable knowledge for robots. arXiv preprint arXiv:2312.02419, 2023

    2023 arXiv

  12. [20]

    Vineet Bhat, Ali Umut Kaypak, Prashanth Krishnamurthy, and et. al. Grounding llms for robot task planning using closed-loop state feedback. arXiv preprint arXiv:2402.08546 , 2024

    2024

  13. [21]

    Achiam Josh, Adler Steven, Agarwal Sandhini, and et. al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 , 2023

    2023 arXiv

  14. [22]

    Wang Xuezhi, Wei Jason, Schuurmans Dale, and et. al. Self- consistency improves chain of thought reasoning in language models. arXiv preprint arXiv:2203.11171 , 2022

    2022 arXiv

  15. [23]

    Madaan Aman, Tandon Niket, Gupta Prakhar, and et. al. Self- refine: Iterative refinement with self-feedback. Advances in Neural Information Processing Systems , 36, 2024

    2024

  16. [24]

    Chaffin Antoine, Claveau Vincent, Kijak Ewa, and et. al. Ppl- mcts: Constrained textual generation through discriminator-guided mcts decoding. arXiv preprint arXiv:2109.13582 , 2021

    2021 arXiv

  17. [25]

    Gu Yu, Deng Xiang, Su Yu, and et. al. Don’t generate, discriminate: A proposal for grounding language models to real-world environments. arXiv preprint arXiv:2212.09736 , 2022

    2022 arXiv

  18. [26]

    Freda Shi, Xinyun Chen, Kanishka Misra, and et. al. Large language models can be easily distracted by irrelevant context. In International Conference on Machine Learning , pages 31210–31227. PMLR, 2023

    2023

  19. [27]

    The pitfalls of next-token prediction

    Gregor Bachmann and Vaishnavh Nagarajan. The pitfalls of next-token prediction. arXiv preprint arXiv:2403.06963 , 2024

    2024 arXiv

  20. [28]

    Puig Xavier, Ra Kevin, Boben Marko, and et. al. Virtualhome: Simulating household activities via programs. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 8494–8502, 2018

    2018

  21. [29]

    Choi Jae-Woo, Yoon Youngwoo, Ong Hyobin, and et. al. Lota-bench: Benchmarking language-oriented task planners for embodied agents. arXiv preprint arXiv:2402.08178 , 2024

    2024 arXiv

  22. [30]

    Singh Ishika, Blukis Valts, Mousavian Arsalan, and et. al. Progprompt: Generating situated robot task plans using large language models. In 2023 IEEE International Conference on Robotics and Automation (ICRA), pages 11523–11530. IEEE, 2023

    2023

  23. [31]

    Li Manling, Zhao Shiyu, Wang Qineng, and et. al. Embodied agent interface: Benchmarking llms for embodied decision making. arXiv preprint arXiv:2410.07166, 2024

    2024 arXiv

  24. [32]

    Chen Junting, Mu Yao, Yu Qiaojun, and et. al. Roboscript: Code generation for free-form manipulation tasks across real and simulation. arXiv preprint arXiv:2402.14623 , 2024

    2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.