Pith. sign in

REVIEW 3 major objections 4 minor 26 references

Internalizing Academic Writing Workflows for Introduction Generation via Struct-Aware Policy Learning

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A single RL-trained policy can generate a structured paper introduction in one pass and rival multi-stage writing workflows and large closed-source models.

desk verdict The method is a plausible extension of stage-token RL for structured writing, but the headline structural metrics are the same functions the policy is trained on, so the claimed edge over closed-source LLMs is only partly supported. read the letter →

arxiv 2608.03138 v1 pith:L6TA2EJF submitted 2026-08-04 cs.CL cs.AI

Meicong Zhang , Tiancheng Su , Jiahao Cheng , Guoxiu He , Xinqi Tao , Dejia Song This is my paper · ORCID
classification cs.CLcs.AI
keywords reinforcementlearningpolicyoptimizationcreditassignmentintroductiongenerationacademicwritingstagetokenssingle-passLLMworkflow
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

Today's best systems for writing a paper introduction split the work into stages—outline, draft, revise—orchestrated through prompts and agent calls. This paper argues that the whole workflow can instead be trained into a single policy that produces an introduction in one pass, with explicit markers separating background, problem, method, and contribution sections and their outlines. The key move is a reward design that judges each stage separately, so a good background is not punished for a weak method paragraph; a global alignment signal still keeps the whole introduction coherent. The authors report that this one-pass model beats workflow-based pipelines on automated metrics, lowers inference tokens by roughly 74%, transfers to a new domain, and at 32B scale is judged about as good as GPT-5.1 by human readers. If right, the result suggests that structured academic writing does not need external orchestration once the policy itself has internalized the workflow.

What carries the argument

The central mechanism is the Struct-aware Relative Advantage (SRA) estimator. Instead of assigning one scalar reward to the entire generated introduction (as in GRPO/PPO), SRA computes a normalized advantage per stage: local advantages compare the same stage across sampled trajectories, a global advantage compares whole introductions, and the two are fused with weight λ. A refinement-gap term subtracts a penalty from draft-stage tokens when the revision version's stage reward is much higher, which pushes the first-pass policy to behave like its own reviser. These stage-level advantages are projected back to the tokens inside the corresponding stage boundaries, so the gradient can reward one

What would settle it

Take a set of introductions with non-canonical rhetoric (theory papers or surveys that skip an explicit contributions section), have human experts judge quality, and compare human rankings to StructPO's structure-score rankings. If introductions the paper's classifier calls structurally correct are systematically rated worse by experts—or if a workflow baseline wins human preference while losing on the automated metrics—the fixed decomposition, not the writing capability, would be carrying the result.

Watch

Extended reading notes

Core claim

StructPO treats an introduction as eight structural units—outline and content for each of four rhetorical sections (background, problem, method, contributions)—generated between explicit stage tokens in a single pass. Training uses draft–revision trajectories: the model first drafts, then revises. Rewards are computed per stage: outlines are scored by bidirectional semantic matching against reference outlines plus a length penalty; content is scored by embedding similarity, a trained classifier's estimate of whether each sentence belongs to its assigned rhetorical section, and Gaussian length regularization. Struct-aware Relative Advantage (SRA) then normalizes these stage rewards within a g

Load-bearing premise

The gains rest on the fixed GPT-4o segmentation of each introduction into the four rhetorical sections being a correct, sufficiently rich ground truth; if that decomposition is biased, the reward signal and the automated metrics are biased in the same direction with no external check.

Editorial extensions

If this is right

  • A single-pass policy with stage tokens can replace handcrafted multi-stage prompts and agent workflows for introduction generation, avoiding cross-stage drift.
  • Distilling revision into the first pass cuts deployment cost: overhead tokens drop by about 74% compared with a workflow baseline while retaining fuller content than SFT stage-token models.
  • Stage-level credit assignment makes policy optimization stable for long-form writing: the ablation shows standard GRPO degrades factual consistency on the 7B backbone, while StructPO does not.
  • The learned rhetorical structure transfers out of domain: an ACL-trained checkpoint applied to CVPR papers keeps structural control and beats GPT-5.1 on structure and section alignment.
  • Scaling the trained policy to 32B parameters brings it to parity with strong closed-source models in blind human evaluation (53.3% win rate), suggesting policy structure and backbone scale are complementary.

Reading between the lines

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

  • If the fixed four-section template is relaxed, the same SRA machinery could in principle learn section count and ordering as part of the policy, replacing the hand-set eight-stage grammar.
  • The refinement-penalty trick suggests a general recipe for 'amortized revision' in any long-form generation task: train the first pass by penalizing exactly the drafts that a later pass would fix, without running the reviser at inference.
  • Because the structure reward relies on a classifier, a natural stress test is to measure reward hacking over longer training: whether structure scores keep rising while human judges see repetitive or template-following text.
  • For neighboring tasks like related-work sections or survey generation, the same stage-token plus dense-advantage design could internalize their multi-stage pipelines, shrinking the cost of agentic writing tools generally.
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

3 major / 4 minor

Summary. The paper proposes StructPO, a structure-aware policy optimization framework for generating academic paper introductions in a single pass using explicit stage tokens. The method decomposes an introduction into eight units (four rhetorical sections × outline/content), computes stage-level rewards for semantic similarity, length, and structural rationality, and uses a Struct-aware Relative Advantage (SRA) estimator with a refinement-guided penalty to train a policy. The authors evaluate on 1,176 ACL 2025 papers and 141 CVPR papers, comparing against prompt-based, workflow-based, and closed-source LLM baselines. They report that StructPO improves semantic alignment, structural rationality, and inference efficiency, and that a Qwen3-32B variant is competitive with GPT-5.1 in a small blind human evaluation.

Significance. If the results held, the paper would make a useful contribution: it formulates structured introduction generation as a single-pass policy optimization problem, provides a concrete dense-reward design, and demonstrates a substantial reduction in inference overhead relative to multi-stage workflows. The paper also includes detailed training configurations, dataset splits, and reproduction details (e.g., Verl, SGLang, hyperparameters), which are valuable strengths. The central limitation is that the automated evaluation metrics are essentially the same functions as the training rewards, so the headline automated gains over closed-source models are partly by construction. The only independent human evidence is a small, non-significant win-rate comparison against GPT-5.1, and it does not address the structural-advantage claim. The paper's own conclusion wisely notes the fixed eight-stage template limitation, but this does not resolve the metric-circularity concern. With an external evaluation of rhetorical structure and appropriate statistical reporting, the central claims could be supported.

major comments (3)
  1. [§Evaluation Metrics / §Reward Design (Eq. 4)] The automated metrics Sem., Sec., Len., and Struc. are, by the paper's own equations, the quantities the policy is trained to maximize. Sem. corresponds to the embedding similarity r_sim used in content rewards, Sec. corresponds to the section-level similarity derived from the GPT-4o decomposition (the same supervision used for outline rewards r_match), Len. corresponds to the Gaussian/linear length regularization r_gau/r_lin, and Struc. corresponds to r_struct computed by the DeBERTa classifier. Eq. (4) and the global reward R_global therefore train against these exact functions. Tables 1 and 2 then report gains on these same functions as 'Auto Overall.' The paper notes that Soundness, Presentation, and Contribution are not used in training, but these are LLM-judge metrics, not independent human ground truth. To support the structural-advantage claim, the authors should add an external
  2. [§Experiments / Appendix 'Prompt for Structural Decomposition'] The ground truth for both reward computation and evaluation is a single GPT-4o decomposition with a fixed prompt. The paper states: 'Stage-level supervision is obtained by using GPT-4o with a fixed decomposition prompt to segment each introduction' and 'The resulting section-content and section-outline pairs are used for supervised training, reward computation and evaluation.' A systematic bias in GPT-4o's segmentation (e.g., misclassifying method sentences as background, or over-partitioning contributions) will be learned by the reward model and then reported as high Sec./Struc. The reported 94.76% classifier accuracy is on the same GPT-4o label distribution, so it does not break the circle. An independent human-annotated segmentation (with inter-annotator agreement) is needed before one can conclude that StructPO produces objectively more rational rhetorical structure.
  3. [§Human Evaluation / Table 4] The only human evaluation is 30 papers × 3 annotators, with no significance testing. The win rate is 53.3% (16/30) and vote share 55.6% (50/90); a two-sided binomial test of 50/90 gives p ≈ 0.36, so the result is not statistically distinguishable from chance. The abstract's claim that StructPO 'remains competitive with GPT-5.1' is appropriately cautious, but the text later says StructPO 'can be preferred over GPT-5.1 in human evaluation,' which is stronger than the data support. Also, this evaluation only compares against GPT-5.1; it does not test the claimed structural advantage over workflow baselines. Please report a proper significance test and, ideally, additional annotation sample size.
minor comments (4)
  1. [Table 3 caption / 'Overall' definition] The ablation table's 'Overall' is defined as a rescaled composite of four automated and two judge metrics, which differs from 'Auto Overall' in Table 2. Although the caption explains this, the shared term 'Overall' is confusing and should be renamed (e.g., 'Composite') to avoid misleading comparisons across tables.
  2. [Fig. 3 caption] The vertical axis mixes 0–100 normalized scores with 1–5 judge scores. The parenthetical values clarify the mapping, but a dual-axis or separate panels would improve readability.
  3. [Eq. (6)] The variance expression Var(g) ∝ Var(r_k) + Σ_{j≠k} Var(r_j) is simply the variance of a sum under independence assumptions. It is not incorrect, but the notation overstates the contribution; the more relevant point is the attribution noise, not the variance decomposition itself.
  4. [Conclusion / fixed template limitation] The paper honestly states that StructPO relies on a fixed eight-stage template and may require adaptation for theoretical, survey, or non-standard papers. This limitation should be discussed earlier, since it qualifies the 'generalization' claim and the CVPR transfer result.

Circularity Check

2 steps flagged · score 6.0 of 10

The headline structural/section metrics are the same GPT-4o-derived reward functions used to train StructPO; only the judge and small human evaluations are independent signals.

  1. fitted input called prediction [Methodology 'Reward Design' (Eq. 4); Experiments 'Evaluation Metrics'; Appendix 'Structural Rationality Illustration']
    "To enforce rhetorical purity, we train a DeBERTa-v3 classifier ... For stage k, the structural rationality score is rstruct(wk) = 1−Cmis/Ctot ... The final local reward depends on the modality of stage k: rk = ... ω1rsim(wk)rstruct(wk)+ω2rgau_len(wk) ... Automated metrics include ... structure score (Struc.) ... The Structural Rationality Score for wk is then computed as rstruct(wk) = 1−Cmis/Ctot."

    The Struc. metric reported in Tables 1–3 is exactly r_struct, the same quantity that appears in the content-stage training reward (Eq. 4). The DeBERTa classifier producing r_struct is itself trained on GPT-4o rhetorical annotations (Appendix). StructPO is therefore explicitly optimized to maximize the function later reported as its Structure Score; comparing that score with baselines that were not trained on this reward shows fit to the training objective, not independently established rhetorical quality.

  2. fitted input called prediction [Experiments 'Dataset Splits and Construction'; Appendix 'Additional Experimental Details']
    "Stage-level supervision is obtained by using GPT-4o ... with a fixed decomposition prompt to segment each introduction into background, problem, method, and contributions sections, together with corresponding outlines. ... The decomposition is performed with GPT-4o ... The resulting section-content and section-outline pairs are used for supervised training, reward computation and evaluation."

    The same GPT-4o-decomposed section references are used as targets for the r_sim/r_match rewards (Eqs. 3–4) and as the basis for the Sec. automated metric ('section-level similarity'). Because the metric is computed against the exact segmentation the policy was rewarded to match, the Sec. gains in Tables 1–2 are partly by construction and do not provide an external check on the quality or correctness of the rhetorical decomposition.

full rationale

StructPO's central automated evidence for structural improvement is not independent of its training signal. The paper explicitly says the GPT-4o section-content/outline pairs are used 'for supervised training, reward computation and evaluation,' and the Structure Score is literally the r_struct term in the training reward (Eq. 4), computed by a DeBERTa classifier trained on GPT-4o labels. Section Similarity is likewise computed against the same GPT-4o-decomposed references used in the r_sim/r_match rewards. Thus the large StructPO-vs-baseline gaps on Sec./Struc. largely reflect optimization against the evaluation function itself, not an external measure of rhetorical quality. The LLM-judge metrics (AWQ, SFC, Soundness, Presentation, Contribution) are not used in training and provide some independent signal, but they are LLM-as-judge, not human ground truth. The only human evaluation is 30 papers × 3 annotators against GPT-5.1, used for the competitiveness claim rather than for validating structural rationality. The same-author STIG baseline is cited but is not load-bearing for the main derivation; the circularity is the reward/evaluation identity. Overall, this is partial circularity: the headline structural claims are partly by construction, while the judge and human evaluations retain independent content.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a particular definition of writing quality encoded in GPT-4o decompositions, embedding similarity, and a section classifier. Together these define the optimization target and the metric, which creates the reward-evaluation circularity noted in the report.

free parameters (6)
  • lambda (global advantage weight) = 0.3
    Chosen in Experiments to weight global vs local advantage in Eq. (9); no sensitivity analysis reported.
  • eta (refinement penalty coefficient) = not reported
    Introduced in Eq. (11); no value or sweep is given.
  • omega_1, omega_2 (content reward weights) = not reported
    Weights in Eq. (4) for semantic/structural and length rewards in content stages; values not specified.
  • delta (Gaussian length bandwidth) = 6 section, 12 global
    Set manually in the reward design; controls how tightly length is enforced.
  • per-stage target lengths L*_k = not reported, presumably reference-derived
    The Gaussian length reward requires target lengths per stage; the values or derivation are not given.
  • rmax_k normalization bounds = not reported
    Eq. (7) normalizes rewards by theoretical upper bounds, but the bounds are not derived or listed.
assumptions (6)
  • domain assumption An introduction can be represented exactly by four rhetorical sections (background, problem, method, contributions), each with outline and content modalities.
    Defines the 8-stage grammar in Eq. (1) and all rewards; the paper acknowledges non-standard papers are out of scope.
  • domain assumption GPT-4o's decomposition of original introductions into sections and outlines is a reliable ground truth.
    Used for SFT targets, reward references, and evaluation; accuracy of this decomposition is not independently validated.
  • domain assumption Cosine similarity in Qwen-Embedding space is a valid proxy for semantic alignment of generated and reference text.
    rsim and the Sem./Sec. metrics rely on this; no correlation with human judgment is shown.
  • domain assumption The DeBERTa-v3 classifier trained on GPT-4o labels accurately identifies rhetorical sections.
    The appendix reports 94.76% accuracy on a held-out set, but that held-out set shares the GPT-4o labeling schema used for rewards.
  • domain assumption The draft-revision penalty in Eq. (11) is a stable and effective way to distill revision behavior.
    No formal convergence argument; ablation support only on two backbones.
  • standard math Importance sampling with clipping and KL regularization gives unbiased policy updates.
    Standard PPO/GRPO machinery cited by the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Internalizing Academic Writing Workflows for Introduction Generation via Struct-Aware Policy Learning." pith.science (2026). https://pith.science/paper/L6TA2EJF

@misc{pith2026260803138,
  author       = {Pith},
  title        = {Pith review of: Internalizing Academic Writing Workflows for Introduction Generation via Struct-Aware Policy Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L6TA2EJF}},
  note         = {Machine review of arXiv:2608.03138}
}
read the original abstract

Generating a rigorous paper introduction with large language models (LLMs) remains challenging, since it requires coordinating background, gap identification, method and contribution within a coherent narrative. Existing solutions externalize this process as multi-stage prompts or agent workflows which are expensive and vulnerable to cross-stage drift. We propose StructPO, a struct-aware policy learning framework that internalizes the entire multi-stage writing workflow into a single-pass policy controlled by explicit stage tokens. StructPO introduces struct-aware credit assignment to decouple local stage quality from global coherence and refinement-guided optimization to internalize revision behavior into the first-pass policy. Experiments show that StructPO improves semantic alignment, structural rationality and inference efficiency over workflow-based baselines, generalizes to out-of-domain settings, and remains competitive with GPT-5.1 in human evaluation when scaled to Qwen3-32B. These results show that internalizing academic writing workflows through fine-grained policy optimization offers a viable alternative to costly external orchestration.

Figures

Figures reproduced from arXiv: 2608.03138 by the authors.

Figure 1
Figure 1. Overview of the StructPO training pipeline. Given paper metadata, the policy generates draft–revision trajectories with [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Inference-token comparison across generation paradigms. StructPO achieves a favorable Pareto balance between [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. Zero-shot transfer results from ACL to 141 unseen CVPR papers. StructPO variants preserve strong structural control [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Illustration of the calculation of the structure score. [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison among AutoSurvey, STIG and StructPO on a real ACL test case (Huang et al. 2025). Blue [PITH_FULL_IMAGE:figures/full_fig_p018_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 21 canonical work pages

  1. [4]

    Lu, C.; Lu, C.; Lange, R

    A Vision for Auto Research with LLM Agents.CoRR, abs/2504.18765. Lu, C.; Lu, C.; Lange, R. T.; Foerster, J. N.; Clune, J.; and Ha, D. 2024. The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery.CoRR, abs/2408.06292. Ouyang, L.; Wu, J.; Jiang, X.; Almeida, D.; Wainwright, C.; Mishkin,P.;Zhang,C.;Agarwal,S.;Slama,K.;Ray,A.;etal

    arXiv 2024

  2. [7]

    theSpeechFakedatasetwasdeveloped

    Leveraging large language models for literature review tasks-a case study using chatgpt. InInternational Confer- ence on Advanced Research in Technologies, Information, Innovation and Sustainability, 313–323. Springer. Structural Rationality Illustration Auxiliaryclassifier.TheStructuralRationalityScorerelies on an auxiliary classifier that predicts the i...

    work page 2021

  3. [21]

    may suggest

    Academic Vocabulary & Formal Tone 16- Uses precise , formal terminology appropriate for scholarly writing 17- Avoids colloquial expressions , slang , or overly casual language 18- Employs appropriate hedging ( e . g . , " may suggest " , " appears to " , " potentially ") 19- Uses domain - specific technical terms correctly 20- Avoids vague expressions ( e...

    work page

  4. [41]

    Background : Basic background and significance of the research field 5- Number of points : 2 -4

    work page

  5. [62]

    Problem and Limitations of Existing Methods : Current issues , challenges and limitations of existing methods 7- Number of points : 2 -6

    work page

  6. [83]

    Brief Method Overview and Summary of Main Results : Overview of the proposed method , main experimental results and findings 9- Number of points : 4 -8

    work page

  7. [104]

    sections

    Our Contributions : Main contributions and innovations of the paper 11- Number of points : 2 -3 12 13Please output in the following JSON format , including outline points and paragraphs classified by section : 14 15{ 16" sections ": { 17" Background ": " Combine all paragraphs and sentences belonging to the background section " , 18" Problem and Limitatio...

    work page

  8. [151]

    ( The content text is about 90 words )

    Background : Provide the research background and the current status of the field . ( The content text is about 90 words )

    work page

Show all 26 references
  1. [162]

    ( The content text is about 180 words )

    Problem and Limitations of Existing Methods : Describe the research problem and the limitations of existing methods . ( The content text is about 180 words )

  2. [173]

    ( The content text is about 230 words )

    Brief Method Overview and Summary of Main Results : Briefly introduce the proposed method and summarize the main results . ( The content text is about 230 words )

  3. [184]

    [1]" , "( Author , Year )

    Our Contributions : Summarize the contributions of this paper . ( The content text is about 70 words ) 19 20Please write introduction with the following writing format and use < STAGE > and <END > markers to represent writing stages : 21< STAGE0 > Outline for Background : 22( ...

  4. [242]

    Section 3 presents ... , Section 4 describes ... , Section 5 shows

    Logical Structure & Argumentation 25- Clear problem statement and research motivation 26- Logical progression : Background - Research Gap - Proposed Solution - Contributions 27- Well - organized paragraphs with clear topic sentences 28- Smooth transitions between ideas and sec...

  5. [341]

    Analyze the content coherently and categorize it into the appropriate sections , strictly controlling the number of points for each section

  6. [352]

    There should be no section 1 , section 2 , section 1 sequences

    When assigning sections , ensure continuity ; for example , Background must be at the beginning of the article and if there is an Our Contributions section , it must be at the end . There should be no section 1 , section 2 , section 1 sequences

  7. [363]

    Some papers may not have a section similar to Our Contributions ; if so , generate an empty Our Contributions field

  8. [374]

    First , divide the sections , then perform an outline analysis to identify key points

  9. [385]

    this " or

    Do not use demonstrative pronouns like " this " or " the model " in the key points ; use specific names if available . Inference and Refinement Prompts The following prompts are used for inference and refinement. 1As an academic writing expert who has completed a research proj...

  10. [473]

    We introduce X , comprising 3 M samples across 46 languages -10 times larger than prior benchmarks

    Content Proportion & Emphasis 48- Background : Should be CONCISE ( ideally 15 -25% of introduction ) , NOT a textbook tutorial 49- Research Gap : Clearly and specifically articulated with concrete examples or evidence 50- Proposed Solution : Should be prominent , with specific...

  11. [644]

    as shown in Figure X

    Sentence Quality & Readability 65- Varied sentence structures ( not repetitive patterns ) 66- Appropriate sentence length ( not too long or choppy ) 67- Clear and unambiguous expressions 68- Proper grammar and syntax 69- No awkward phrasing or unclear constructions 70- PENALIZ...

  12. [735]

    With the rapid advancement of X

    Introduction - Specific Requirements 74- Opening : Engaging first paragraph that establishes importance and relevance 75- BAD Opening : " With the rapid advancement of X ..." / " In recent years , X has attracted ..." 76- GOOD Opening : Start with a concrete problem , striking...

  13. [816]

    contributions can be summarized as follows

    Professional Academic Style 82- Appropriate use of active / passive voice 83- Objective and impersonal tone where appropriate 84- Concise writing without unnecessary redundancy 85- PENALIZE : Redundant phrases like " contributions can be summarized as follows " ( just say " Ou...

    2025

  14. [1478]

    He, P.; Gao, J.; and Chen, W

    Online: Association for Computational Linguistics. He, P.; Gao, J.; and Chen, W. 2021. DeBERTaV3: Improving DeBERTausingELECTRA-StylePre-TrainingwithGradient- Disentangled Embedding Sharing. arXiv:2111.09543. He, P.; Liu, X.; Gao, J.; and Chen, W. 2021. DEBERTA: DECODING-ENHAN...

    2021 arXiv

  15. [2022]

    Schmidgall, S.; Su, Y.; Wang, Z.; Sun, X.; Wu, J.; Yu, X.; Liu, J.; Moor, M.; Liu, Z.; and Barsoum, E

    Training language models to follow instructions with human feedback.Advances in neural information processing systems, 35: 27730–27744. Schmidgall, S.; Su, Y.; Wang, Z.; Sun, X.; Wu, J.; Yu, X.; Liu, J.; Moor, M.; Liu, Z.; and Barsoum, E. 2025. Agent Laboratory: Using LLM Agen...

    2025 arXiv

  16. [2024]

    Zimmermann, R.; Staab, M.; Nasseri, M.; and Brandtner, P

    Sglang: Efficient execution of structured language model programs.Advances in neural information processing systems, 37: 62557–62583. Zimmermann, R.; Staab, M.; Nasseri, M.; and Brandtner, P

  17. [2025]

    Garikaparthi, A.; Patwardhan, M.; Kanade, A

    Let’sUseChatGPTToWriteOurPaper!Benchmarking LLMs To Write the Introduction of a Research Paper.arXiv preprint arXiv:2508.14273. Garikaparthi, A.; Patwardhan, M.; Kanade, A. S.; Hassan, A.; Vig, L.; and Cohan, A. 2025. MIR: Methodology Inspira- tion Retrieval for Scientific Res...

    2025 arXiv

  18. [2026]

    arXiv:2601.05688

    SketchVL: Policy Optimization via Fine-Grained Credit Assignment for Chart Understanding and More. arXiv:2601.05688. Huang, W.; Gu, Y.; Wang, Z.; Zhu, H.; and Qian, Y. 2025. SpeechFake: A Large-Scale Multilingual Speech Deepfake Dataset Incorporating Cutting-Edge Generation Me...

    2025

Pith tools

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