Pith. sign in

REVIEW 3 major objections 6 minor 37 references

Interaction Scaling: Grounding the Third Axis of Test-Time Compute

T0 review · 3 major / 6 minor · reviewed 2026-07-14 · grok-4.5

Pith's one-line read Interaction with grounded instruments is a third axis of test-time compute that breaks the ceiling reasoning and sampling hit.

desk verdict Code axis is clean and strong: grounded interaction hits 100% with zero variance where reasoning and oracle best-of-N plateau; visual half is real but partly optimizes its own scorer, so the full two-sided claim needs the human check they flag. read the letter →

arxiv 2607.11598 v1 pith:JZMXIHO5 submitted 2026-07-13 cs.AI

classification cs.AI
keywords test-timecomputeinteractionscalinggroundingproposer-reviewerharnessexecutionfeedbacklayoutgeometrymodel-as-judgecoverageprinciple
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 argues that test-time compute has three axes, not two. Reasoning longer and sampling more attempts are both internal: every extra token still comes from the same frozen weights and the same prompt, so neither can tell the model anything it does not already know. Interaction—proposing an artifact, letting an external instrument observe how it actually behaves, then revising—imports a real observation each cycle and can keep improving past that internal ceiling. The single governing variable is grounding, and it must hold on both sides of the loop: the feedback that drives revision must come from an instrument that actually sees the flaw, and so must the metric that scores the result. On hard coding tasks the interaction harness reaches a perfect pass rate with no variance while reasoning and even oracle best-of-N plateau; on visual artifacts the usual screenshot judge is structurally blind, and only a geometry-measuring tool reveals the large defect reductions the loop actually achieves.

What carries the argument

Grounding on both sides of the interaction loop, plus the coverage principle: feedback is grounded only when an external instrument observes the artifact’s real form or behavior, and it helps exactly as far as that instrument can observe the defects present. The same observation must also serve as the evaluation metric, or the gain remains invisible.

What would settle it

A human preference study on the same figure, slide, web, and animation suites showing that lower measured DOM-defect counts do not predict human preference, or a matched-budget experiment in which oracle best-of-N equals or exceeds the grounded interaction harness once human raters replace the instrument score.

Watch

Extended reading notes

Core claim

At a fixed token budget on hard coding tasks, reasoning-only and best-of-N sampling (even with an oracle verifier) both saturate, while every strategy that iterates on grounded execution feedback keeps climbing; a proposer–reviewer harness reaches 100% pass rate with zero seed variance across three model families. On rendered visual artifacts the default vision-language model judge rates 14 of 15 visibly broken figures “perfect” because screenshots hide the flaws; a deterministic layout instrument instead shows the same loop removing 40–74% of defects, and that same VLM used as reviewer makes slides worse where the measuring tool repairs them.

Load-bearing premise

The paper treats reductions in its deterministic DOM-geometry defect counts—using fixed pixel thresholds for overlap, overflow, and misalignment—as faithful measures of quality a human would prefer.

Editorial extensions

If this is right

  • Reasoning and sampling alone cannot produce candidates outside the high-probability region of the model’s own distribution, even with perfect selection.
  • A frozen frontier model wrapped in a proposer–reviewer harness that feeds instrument observations reaches higher quality at lower token cost and with zero run-to-run variance.
  • For layout-sensitive visual artifacts, screenshots must be replaced by direct geometry measurement on both the reviewer and the scorer.
  • Using an ungrounded model as reviewer can actively regress quality on defects its channel cannot see.
  • Distilling teacher interaction trajectories into a small student recovers a substantial fraction of teacher quality at much lower deployment cost, provided sampling variance is retained.

Reading between the lines

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

  • Results in self-improving visual generation that rely solely on VLM-on-screenshot judges may systematically overstate quality or miss real layout failures.
  • The two-sided grounding requirement extends naturally to audio, 3D, tabular, and interactive UI artifacts where model judges are similarly blind to physical properties.
  • Industrial coding harness gains already seen in large deployments are predicted once the instrument’s coverage is matched to the defect class.
  • Post-training that collapses output variance can trade away the very resource best-of-N sampling converts into solved tasks.
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 / 6 minor

Summary. The paper argues that test-time compute has a third axis, interaction, distinct from internal reasoning and sampling: a frozen model proposes an artifact, an external instrument observes its actual behavior, and the model revises. The governing variable is two-sided grounding—both the feedback that drives revision and the metric that scores the result must come from an instrument that observes the relevant defect (the coverage principle). On hard coding tasks at matched token budgets, reasoning-only and oracle best-of-N plateau while a proposer–reviewer harness reaches 100% pass rate with zero seed variance across three model families and a held-out suite. On rendered visual artifacts, a screenshot VLM rates most broken figures perfect and can regress slide layouts as a reviewer, whereas a deterministic DOM-geometry instrument reports 40–74% defect reductions across four modalities. Two modalities (video, deep research) are reported as honest negatives with no single-shot headroom.

Significance. If the results hold, the paper cleanly separates interaction from reasoning and sampling under a matched budget, with an information-theoretic motivation (Appendix A) and a falsifiable coverage principle that organizes the ablations. The code evidence is unusually strong for this literature: fixed budgets, oracle best-of-N as an internal ceiling, three seeds, cross-family replication (Sonnet 4, Qwen3-235B, GPT-5), held-out recovery with no regressions, sign tests and bootstrap CIs, and released code. The structural-blindness critique of screenshot VLM judges is practically important and well illustrated. The distillation section’s caution that RFT can spend the variance best-of-N needs is a useful secondary finding. These strengths make the work a serious contribution to test-time compute, contingent on tightening the visual evaluation half of the two-sided claim.

major comments (3)
  1. [Section 6; Section 9; Figure 12; Table 6] Section 6 and Limitations (Section 9): the visual half of the central claim (Prediction 3; Figs. 10–12; Table 6) rests on DOM-geometry defect counts that both drive revision and score the result, with keep-best guaranteeing non-worsening. Thresholds (≥6 px text overlap, >16 px document overflow, box-group misalignment rules) are free parameters. The authors correctly flag that a quantitative human-preference study is future work, but without it the reported 40–74% reductions and the claim that a VLM reviewer regresses quality are only decisive under an unvalidated metric. This is load-bearing for the two-sided grounding thesis. A modest preference study (or at least a blinded human defect audit on a stratified sample of single-shot vs. reviewed pairs) is needed before the visual results can carry equal weight with the code results.
  2. [Section 3; Table 6] Section 3 and Table 6: the geometry instrument’s defect definitions and thresholds are not accompanied by a sensitivity analysis. Because the same instrument is the sole reported scorer, it is unclear how much of the effect size depends on the particular cutoffs versus robust layout failures. Reporting defect reductions under a small grid of thresholds (or an ablated subset of defect classes) would show that the qualitative claim is not an artifact of one hand-chosen rule set.
  3. [Section 4; Figure 10; Table 2; Table 6] Section 4 / Table 2 and the visual suites: several headline numbers rest on modest N (15 hard code tasks; 12 dense slides; 15–20 per visual modality). The code path is partly rescued by three seeds, cross-family replication, and a 32-task held-out suite (Table 8), but the visual effect sizes and the VLM-reviewer regression (Figure 10) would be more convincing with larger or more diverse task sets, or at least explicit power discussion for the paired sign tests. This is secondary to the human-validation issue but still load-bearing for generalizing Prediction 3.
minor comments (6)
  1. [Figure 5; Table 2] Figure 5 and Table 2: the small non-monotonic dip in best-of-N from B=1K to B=5K is explained as seed noise, but a one-sentence note in the main text (not only the table caption) would prevent misreading of the internal-ceiling plot.
  2. [Section 2.2; Figure 4] Section 2.2 / Figure 4: classifying a screenshot-fed VLM as ungrounded is justified in the text, but the figure’s orange lane could more explicitly mark that the lossy channel, not the pixels alone, is what places it in the ungrounded category.
  3. [Section 2.3; Appendix A] Appendix A: the DPI argument correctly bounds information channels rather than achievable quality; the main text already notes this, but a single forward pointer in Section 2.3 to the two cautions at the end of Appendix A would help readers who stop at the informal one-sentence claim.
  4. [Table 3] Table 3 vs. Table 2: the single-seed reasoning-only numbers and the 3-seed aggregate can be misread as the same run; a clearer cross-reference in the caption of Table 3 would reduce confusion.
  5. [Section 7] Section 7 / Figure 14: the RFT variance finding is interesting but somewhat orthogonal; consider shortening or moving more of the distillation tables to the appendix so the main narrative stays on the three predictions.
  6. Minor polish: arXiv date line and a few long figure captions (e.g., Figure 1) could be tightened; ensure all modality N values in Table 1 match the statistics tables.

Circularity Check

2 steps flagged · score 4.0 of 10

Code path is non-circular (external pytest oracle shared with best-of-N); visual path has real circular content because the same DOM-geometry instrument both drives revision and scores, and keep-best guarantees non-worsening under that metric.

  1. self definitional [Section 3 (Architecture / Budget and protocol) + Section 6 (Is the reduction circular?)]
    "A reviewer (the same model in a diagnostic role) turns that signal into a structured list of defects, and the proposer revises. The loop repeats up to an iteration cap and keeps the best-scoring iteration, so the reviewed result can never score below single-shot under the same metric. ... Is the reduction circular? The instrument both drives the revision and scores it, and the harness keeps the best-scoring iteration, so some reduction is mechanically guaranteed."

    The DOM-geometry instrument is defined as both the feedback channel and the evaluation metric. Combined with keep-best, non-worsening (and some positive reduction) under that metric is true by construction of the harness, not an independent empirical discovery. The paper acknowledges the mechanical guarantee; the remaining claim is only about magnitude and human relevance, which the authors defer to future human-preference work.

  2. fitted input called prediction [Section 3 (Rendered geometry instrument) + Section 6 / Figure 12 / Table 6]
    "We compute, exactly: text-on-text overlap (≥ 6 px), out-of-bounds/clipping, container overflow, document overflow (> 16 px), and box-group misalignment (rows or columns of card-like boxes flagged for unequal size, misaligned far edges, or uneven gutters). ... Grounded geometric feedback removes real layout defects on all four visual modalities. Mean defect reduction ... −74% / −73% / −47% / −40% ... The principal remaining caveat is a quantitative human-preference study confirming that DOM-defect reduction maps onto perceived quality across the full suite."

    The defect counts that are reported as the main visual result (40–74% reductions) are produced by hand-chosen thresholds and alignment rules that the same instrument optimizes. Without an external human-preference or independent quality mapping, the 'prediction' that grounded interaction improves visual quality reduces to optimization of the authors' own defect definition. The ungrounded-VLM ablation (Figure 10) is a useful control under the same metric but does not break the circle relative to human quality.

full rationale

The paper's central thesis is two-sided grounding: interaction escapes the internal ceiling, but only when both feedback and metric are grounded. On code this is cleanly non-circular. The instrument is pytest pass/fail + traceback; the oracle best-of-N baseline selects against the same tests (Section 3, Section 4, Table 2), so the 100% ceiling and the gap past oracle best-of-N are not defined by a model opinion or by a fitted parameter. Cross-family zero-variance 100% and held-out recovery further show the result is not a tautology of the harness. The visual half is different. Section 3 defines a deterministic DOM-geometry instrument (text-on-text overlap ≥6 px, document overflow >16 px, box-group misalignment rules) that both feeds the defect list into revision and is the sole reported scorer for figures, slides, web, and animations. The harness 'keeps the best-scoring iteration, so the reviewed result can never score below single-shot under the same metric' (Section 3). Section 6 itself asks 'Is the reduction circular?' and answers with three mitigations (real defects, non-trivial magnitude, ungrounded-VLM-reviewer ablation that moves the same metric the wrong way), while Section 9 and the Section 6 caveats correctly flag that a quantitative human-preference study remains future work. Those mitigations reduce but do not eliminate the construction: some reduction is mechanically guaranteed by keep-best under the instrument that also drives the loop, and the claim that ungrounded VLM reviewers regress quality (Figure 10) is decisive only under that same unvalidated metric. No self-citation chain or uniqueness theorem is load-bearing; the circularity is the fitted-input / self-definitional pattern on the visual metric. Overall score 4: partial circularity confined to the visual half of Prediction 3 and the two-sided framing; the code axis and the information-theoretic internal-ceiling argument stand independently.

Assumptions & free parameters 4 free parameters · 5 assumptions · 3 invented entities

The central claim rests on standard information-theory background, domain assumptions about what counts as an external observation, hand-chosen geometry thresholds and harness hyperparameters, and conceptual entities (internal ceiling, coverage principle, two-sided grounding) that organize the experiments rather than introduce new physical objects. No fitted physical constants; free parameters are experimental design choices that affect measured effect sizes.

free parameters (4)
  • geometry defect thresholds (overlap ≥6 px, document overflow >16 px, box-group misalignment rules)
    Hand-chosen cutoffs that define what counts as a layout defect and therefore the reported 40–74% reductions; changing them would change measured effect sizes.
  • iteration caps (≤3 geometry, ≤5 code) and keep-best-iteration rule
    Harness hyperparameters that bound revision and mechanically prevent score regression under the same metric.
  • token budget B and proposer/reviewer split
    Fixed budgets (e.g. 1K/5K/20K) and propose-heavy allocation chosen by simplex sweep; pass rates depend on these design choices.
  • design-principle generation prompt and temperature settings
    Shared prompt (proximity, alignment, repetition, contrast) and T=0 (or 0.7 for some replications) set single-shot headroom and variance.
assumptions (5)
  • standard math Data-processing inequality: ungrounded post-processing of model outputs cannot increase information about the correct artifact beyond what is in weights and prompt (Appendix A).
    Used to motivate the internal ceiling; authors note it bounds the information channel, not quality directly, and treat experiments as the evidence.
  • domain assumption An instrument observation E = g(A, world) can carry information about defects not already in (Θ, X), so conditioning generation on E can produce candidates the model would not sample unaided.
    Load-bearing modeling assumption separating external from internal test-time compute (Section 2.3).
  • ad hoc to paper Coverage principle: a feedback channel helps exactly as far as its instrument’s observational reach.
    Stated as the unifying principle of the paper (Section 2.2); predicts linter nulls, VLM failures, and geometry success.
  • ad hoc to paper Screenshot-fed VLM signal is classified as ungrounded because the model’s reading of a lossy view is what enters the loop.
    Taxonomy choice that places VLM-on-screenshot in the ungrounded lane (Figure 4); drives Prediction 3.
  • domain assumption pytest multi-assertion suites and DOM bounding boxes are faithful external ground truth for behavioral correctness and layout quality respectively.
    Assumed throughout Sections 3–6; layout side lacks human-preference confirmation.
invented entities (3)
  • internal ceiling (of reasoning and sampling) independent evidence
    purpose: Name the plateau of internal test-time compute that grounded interaction is claimed to escape.
    Conceptual bound motivated by DPI and measured in Figure 5; not a new physical object, but a paper-specific construct.
  • two-sided grounding (feedback + evaluation) independent evidence
    purpose: Govern when interaction scaling fires and when it is measurable.
    Central organizing variable of the thesis (Figure 1); operationalized via instrument swaps and metric swaps.
  • coverage principle independent evidence
    purpose: Predict which instruments help which defect classes.
    Unifies code and visual controls; tested by linter vs execution and VLM vs geometry swaps.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Interaction Scaling: Grounding the Third Axis of Test-Time Compute." pith.science (2026). https://pith.science/paper/JZMXIHO5

@misc{pith2026260711598,
  author       = {Pith},
  title        = {Pith review of: Interaction Scaling: Grounding the Third Axis of Test-Time Compute},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JZMXIHO5}},
  note         = {Machine review of arXiv:2607.11598}
}
read the original abstract

There are two standard ways to spend more compute at test time: let a model reason longer, or sample more attempts and keep one. Both share a hidden limit: they are internal. Every extra token comes from the same frozen weights and the same prompt, so neither can tell the model anything it does not already know. We study a third way, interaction: the model proposes an artifact, an external instrument observes how it actually behaves, and the model revises. Each cycle imports a real observation, so interaction breaks through the ceiling the other two hit. We argue that a single variable governs this third axis, grounding, and that it must hold on both sides of the loop. The feedback that drives revision must come from an instrument that actually observes the flaw, and so must the metric that scores the result. On hard coding tasks at a fixed token budget, reasoning-only and best-of-N sampling both plateau (the latter even when an oracle picks the best sample), while every interaction strategy keeps improving; our proposer-reviewer harness reaches a perfect 100% pass rate with no run-to-run variance, and the gain holds across three model families. On rendered visual artifacts, the usual judge (a vision-language model, or VLM, reading a screenshot) rates 14 of 15 visibly broken figures "perfect," because the screenshot hides the flaws before the judge can see them. A tool that measures the real layout instead shows the loop removing 40-74% of defects across four modalities; and that same VLM, used as the reviewer, makes slide layouts worse where the measuring tool repairs them. Interaction scaling is real and distinct from reasoning and sampling, but only visible when both the feedback and the metric are grounded.

Figures

Figures reproduced from arXiv: 2607.11598 by the authors.

Figure 1
Figure 1. Grounding must hold on both sides of the interaction loop. One instrument observation is both (1) grounded feedback (the defect list that drives revision and escapes the internal ceiling) and (2) grounded evaluation, the score that makes the gain measurable. The default VLM-on-a-screenshot judge (orange lane) breaks both: the screenshot drops the defects before the model sees them. 1 arXiv:2607.11598v1 [cs.AI] 13 Ju… view at source ↗
Figure 2
Figure 2. Results at a glance: one grounded loop, seven modalities. (a) Execution feedback lifts both code suites to a strict 100% pass rate, recovering every single-shot failure (3-seed means; Section 4). (b) Grounded geometry feedback removes 40–74% of the layout defects a deterministic DOM instrument measures, on all four visual modalities (every reduction statistically decisive; Section 6). (c) The two remaining modalitie… view at source ↗
Figure 3
Figure 3. A concrete example of what the loop fixes. A dense real-paper slide task (the GAN paper), rendered full-page at equal width: single-shot (a) vs. reviewed (b). The deterministic geometry instrument counts 6 defects single-shot and 0 after review (the same 6 → 0 reduction replicates in a second seed), and a screenshot judge sees none of them (Section 6). tool that measures the rendered layout instead reveals large, st… view at source ↗
Figures from the paper (12 more)
Figure 4
Figure 4. Figure 4: The feedback taxonomy. Feedback is either a model’s opinion (ungrounded, top) or an instrument’s observation of the artifact (grounded, bottom); grounded subtypes are named by what the instrument observes, and the coverage axis orders them by reach. The screenshot-fed …
Figure 5
Figure 5. Figure 5: Internal scaling saturates; external scaling does not (Prediction 1). Code pass rate vs. per-task token budget on the 15 hard tasks (3-seed means; full table in [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Execution feedback recovers every first-shot failure on both code suites. Single-shot vs. harness-reviewed pass rate (development suite: 3-seed mean, whisker ±1 SD; deep-spec: sign test p = 0.002). Every failing run is recovered and no passing run regresses; the review…
Figure 7
Figure 7. Figure 7: Same ceiling, different cost and reliability. All three interaction strategies use execution feedback and tie on pass rate within seed noise (sign test p > 0.6); the proposer–reviewer harness converges with ∼28% fewer tokens and is the only variant at 100% across all s…
Figure 8
Figure 8. Figure 8: The execution loop replicates across model families. Single-shot vs. harness-reviewed code pass rate for three families, three seeds each (lifts +33.3/+22.2/+20.0 pp; whiskers ±1 SD across seeds). The reviewed bars carry zero seed variance for all three families. criti…
Figure 9
Figure 9. Figure 9: Swapping only the feedback signal on a fixed code suite: improvement tracks coverage, and cost tracks it decisively. One reviewing pass per configuration. Ungrounded critique (orange), critique + linter (light blue: grounded, but observing only form), and critique + te…
Figure 10
Figure 10. Figure 10: Same tasks, same proposer, one reviewing pass: the arrow’s direction is set by the reviewer’s instrument. Each arrow runs from that configuration’s own single-shot baseline (open circle; the configurations are separate runs, hence differing baselines) to its reviewed …
Figure 11
Figure 11. Figure 11: The default VLM judge is structurally blind to layout defects. (a) On the same 15 single￾shot academic figures, the VLM-on-a-screenshot judge rates 14 “perfect” while the DOM-geometry instrument finds only 3 actually clean; the other 11 are broken in ways cropped off-…
Figure 12
Figure 12. Figure 12: Grounded geometric feedback removes real layout defects on all four visual modalities. Mean defect reduction under one identical configuration (3 seeds; whiskers are paired-bootstrap 95% CIs, all excluding zero; right column counts improved vs. regressed task-runs; al…
Figure 13
Figure 13. Figure 13: An 8B student internalizes much of the teacher’s interaction quality. Distilled student as a fraction of the frontier teacher on the 44-task out-of-distribution suite ( [PITH_FULL_IMAGE:figures/full_fig_p014_13.png]
Figure 14
Figure 14. Figure 14: RFT polishes consistency but spends the variance that sampling needs. Judge-keep vs. number of samples on the 18-task hard held-out suite for the chosen SFT student and the same student after RFT. RFT improves every per-turn consistency metric ( [PITH_FULL_IMAGE:figu…
Figure 15
Figure 15. Figure 15: Pass rate is monotone in the proposer’s budget share. Nine allocations of a fixed 10K￾token budget across proposer / execution / reviewer, on the hard code suite (open markers: a second allocation with the same proposer share but a different execution/review split, sh…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

37 extracted references · 4 linked inside Pith

  1. [1]

    Advances in Neural Information Processing Systems , year =

    Chain-of-Thought Prompting Elicits Reasoning in Large Language Models , author =. Advances in Neural Information Processing Systems , year =

  2. [2]

    International Conference on Learning Representations , year =

    Self-Consistency Improves Chain of Thought Reasoning in Language Models , author =. International Conference on Learning Representations , year =

  3. [3]

    Advances in Neural Information Processing Systems , year =

    Tree of Thoughts: Deliberate Problem Solving with Large Language Models , author =. Advances in Neural Information Processing Systems , year =

  4. [4]

    Yao, Shunyu and Zhao, Jeffrey and Yu, Dian and Du, Nan and Shafran, Izhak and Narasimhan, Karthik and Cao, Yuan , booktitle =

  5. [5]

    Advances in Neural Information Processing Systems , year =

    Reflexion: Language Agents with Verbal Reinforcement Learning , author =. Advances in Neural Information Processing Systems , year =

  6. [6]

    Advances in Neural Information Processing Systems , year =

    Self-Refine: Iterative Refinement with Self-Feedback , author =. Advances in Neural Information Processing Systems , year =

  7. [7]

    NeurIPS Foundation Models for Decision Making Workshop , year =

    Voyager: An Open-Ended Embodied Agent with Large Language Models , author =. NeurIPS Foundation Models for Decision Making Workshop , year =

  8. [8]

    Annual Meeting of the Association for Computational Linguistics (Short Papers) , year =

    Teaching Small Language Models to Reason , author =. Annual Meeting of the Association for Computational Linguistics (Short Papers) , year =

Show all 37 references
  1. [9]

    Annual Meeting of the Association for Computational Linguistics , year =

    Large Language Models Are Reasoning Teachers , author =. Annual Meeting of the Association for Computational Linguistics , year =

  2. [10]

    International Conference on Machine Learning , year =

    Specializing Smaller Language Models towards Multi-Step Reasoning , author =. International Conference on Machine Learning , year =

  3. [11]

    Advances in Neural Information Processing Systems , year =

    Toolformer: Language Models Can Teach Themselves to Use Tools , author =. Advances in Neural Information Processing Systems , year =

  4. [12]

    and Zhang, Tianjun and Wang, Xin and Gonzalez, Joseph E

    Patil, Shishir G. and Zhang, Tianjun and Wang, Xin and Gonzalez, Joseph E. , booktitle =. Gorilla: Large Language Model Connected with Massive

  5. [13]

    2021 , eprint =

    Evaluating Large Language Models Trained on Code , author =. 2021 , eprint =

  6. [14]

    2024 , eprint =

    Large Language Monkeys: Scaling Inference Compute with Repeated Sampling , author =. 2024 , eprint =

  7. [15]

    Understanding the Effects of

    Kirk, Robert and Mediratta, Ishita and Nalmpantis, Christoforos and Luketina, Jelena and Hambro, Eric and Grefenstette, Edward and Raileanu, Roberta , booktitle =. Understanding the Effects of

  8. [16]

    International Conference on Learning Representations , year =

    Does Writing with Language Models Reduce Content Diversity? , author =. International Conference on Learning Representations , year =

  9. [17]

    Snell, Charlie and Lee, Jaehoon and Xu, Kelvin and Kumar, Aviral , booktitle =. Scaling

  10. [18]

    2501.12948 , archivePrefix =

    DeepSeek-AI , year =. 2501.12948 , archivePrefix =

  11. [19]

    Kimi K1.5: Scaling Reinforcement Learning with

    Kimi Team , year =. Kimi K1.5: Scaling Reinforcement Learning with. 2501.12599 , archivePrefix =

  12. [20]

    Advances in Neural Information Processing Systems , year =

    Learning to Summarize with Human Feedback , author =. Advances in Neural Information Processing Systems , year =

  13. [21]

    2023 , eprint =

    Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback , author =. 2023 , eprint =

  14. [22]

    2024 , howpublished =

    Scaling Test-Time Compute with Open Models , author =. 2024 , howpublished =

  15. [23]

    International Conference on Learning Representations , year =

    Large Language Models Cannot Self-Correct Reasoning Yet , author =. International Conference on Learning Representations , year =

  16. [24]

    When Can

    Kamoi, Ryo and Zhang, Yusen and Zhang, Nan and Han, Jiawei and Zhang, Rui , journal =. When Can

  17. [25]

    Gou, Zhibin and Shao, Zhihong and Gong, Yeyun and Shen, Yelong and Yang, Yujiu and Duan, Nan and Chen, Weizhu , booktitle =

  18. [26]

    Single-Agent

    Tran, Dat and Kiela, Douwe , year =. Single-Agent. 2604.02460 , archivePrefix =

  19. [27]

    The Practice and Exploration of

    Hong, Dingkun , year =. The Practice and Exploration of

  20. [28]

    2023 , eprint =

    Improving Factuality and Reasoning in Language Models through Multiagent Debate , author =. 2023 , eprint =

  21. [29]

    Advances in Neural Information Processing Systems , year =

    Thinking vs.\ Doing: Agents that Reason by Scaling Test-Time Interaction , author =. Advances in Neural Information Processing Systems , year =

  22. [30]

    Gehring, Jonas and Zheng, Kunhao and Copet, Jade and Mella, Vegard and Cohen, Taco and Synnaeve, Gabriel , booktitle =

  23. [31]

    2006 , publisher =

    Elements of Information Theory , author =. 2006 , publisher =

  24. [32]

    2509.22644 , archivePrefix =

    Lu, Zimu and Ren, Houxing and Yang, Yunqiao and Wang, Ke and Zong, Zhuofan and Pan, Junting and Zhan, Mingjie and Li, Hongsheng , year =. 2509.22644 , archivePrefix =

  25. [33]

    Iterative Agent Decoding for Detecting Compounding Errors in

    Ruan, Yangjun and Briakou, Eleftheria and Han, Charles Xie and Chen, Yu and Jiao, Yufan and others , year =. Iterative Agent Decoding for Detecting Compounding Errors in. 2504.01931 , archivePrefix =

  26. [34]

    and Zhang, Hao and Gonzalez, Joseph E

    Zheng, Lianmin and Chiang, Wei-Lin and Sheng, Ying and Zhuang, Siyuan and Wu, Zhanghao and Zhuang, Yonghao and Lin, Zi and Li, Zhuohan and Li, Dacheng and Xing, Eric P. and Zhang, Hao and Gonzalez, Joseph E. and Stoica, Ion , booktitle =. Judging

  27. [35]

    2026 , howpublished =

    Loopcraft: The Art of Stacking Loops , author =. 2026 , howpublished =

  28. [36]

    2026 , howpublished =

    The Art of Loop Engineering , author =. 2026 , howpublished =

  29. [37]

    2026 , howpublished =

Pith tools

Reviewed July 14, 2026 · model on record in the stance chip above.