Paper Citation Record · LEDGER

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B

As of 10 August 2026, this Paper Citation Record lists 37 of 37 outbound references and 0 inbound Pith citation observations for arXiv:2607.28576.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2607.28576 v1

Coverage vector

measured 37 of 37 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-07-31T03:19:01.477082Z

measured 37 of 37 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-10T06:31:04.303077+00:00

measured 0 of 0 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links

measured 0 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: cited_works

Reference resolution

37 of 37 outbound references displayed

verified exact0
verified fuzzy0
unresolved37
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation ce7f7f41-d48b-4e80-8c9d-4f2eacf33c24 · outbound

This paper cites Graph of Thoughts: Solving Elaborate Problems with Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Graph of Thoughts: Solving Elaborate Problems with Large Language Models

Reference 1

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.572680Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e21a5e66-a140-435b-8796-cf71034a8cf9 · outbound

This paper cites Large Language Monkeys: Scaling Inference Compute with Repeated Sampling.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Large Language Monkeys: Scaling Inference Compute with Repeated Sampling

Reference 2

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.616128Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f6c28381-4b71-42ca-9d6d-8a16ce472e79 · outbound

This paper cites Debate or vote: Which yields better decisions in multi-agent large lan- guage models? InAdvances in Neural Information Processing Systems (NeurIPS), Spotlight,.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Debate or vote: Which yields better decisions in multi-agent large lan- guage models? InAdvances in Neural Information Processing Systems (NeurIPS), Spotlight,

Reference 3

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.652283Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 99d3e645-2e73-4760-a8ba-f84dba51c93b · outbound

This paper cites Training Verifiers to Solve Math Word Problems.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Training Verifiers to Solve Math Word Problems

Reference 4

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.690707Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1cacb16a-b6cb-4228-93e2-c3c31f6b526e · outbound

This paper cites Improving Factuality and Reasoning in Language Models through Multiagent Debate.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Improving Factuality and Reasoning in Language Models through Multiagent Debate

Reference 5

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.728716Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 86cb38bb-1d49-4a79-af60-a372adbf204d · outbound

This paper cites Bootstrap methods: Another look at the jackknife.The Annals of Statistics, 7(1):1–26, 1979.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Bootstrap methods: Another look at the jackknife.The Annals of Statistics, 7(1):1–26, 1979

Reference 6

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.769984Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1f741c85-fea5-4a71-a7e4-f59f0155ce51 · outbound

This paper cites llama.cpp, 2026.https://github.com/ ggml-org/llama.cpp.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B llama.cpp, 2026.https://github.com/ ggml-org/llama.cpp

Reference 7

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.809324Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c599eef7-51cb-4aaa-8c3e-763040c9b49d · outbound

This paper cites Measuring Mathematical Problem Solving With the MATH Dataset.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Measuring Mathematical Problem Solving With the MATH Dataset

Reference 8

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.819809Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 34e9b2d7-162d-4ef6-b3ca-5ae0840f0b3d · outbound

This paper cites A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics, 6(2):65–70, 1979.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B A simple sequentially rejective multiple test procedure.Scandinavian Journal of Statistics, 6(2):65–70, 1979

Reference 9

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.861811Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cdb006de-9c05-46a9-85ee-f7926df310b8 · outbound

This paper cites V-STaR: Training Verifiers for Self-Taught Reasoners.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B V-STaR: Training Verifiers for Self-Taught Reasoners

Reference 10

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.902163Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f3ed3f3-df60-41e2-9d5d-44f54f1414e6 · outbound

This paper cites Large Language Models Cannot Self-Correct Reasoning Yet.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Large Language Models Cannot Self-Correct Reasoning Yet

Reference 11

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.943645Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a3fa1e7-b0e4-43ca-862e-76b7460d077c · outbound

This paper cites When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs

Reference 12

Resolution: unresolved; no resolver link, observed 2026-07-31T03:18:59.984759Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d2fa40f4-9416-4a92-9d85-da7ad8d1a133 · outbound

This paper cites Scalable best-of-n selection for large language models via self-certainty.arXiv preprint arXiv:2502.18581, 2025.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Scalable best-of-n selection for large language models via self-certainty.arXiv preprint arXiv:2502.18581, 2025

Reference 13

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.058675Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3158fbd8-1a35-4f98-8b9b-1f9b132e1cef · outbound

This paper cites Large Language Models are Zero-Shot Reasoners.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Large Language Models are Zero-Shot Reasoners

Reference 14

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.116288Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6d307dc6-a7a9-4bd8-a21d-fdc458ca8737 · outbound

This paper cites Let's Verify Step by Step.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Let's Verify Step by Step

Reference 15

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.187681Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 73fc699e-70fd-4600-b3f9-f9ee9b52203e · outbound

This paper cites Self-Refine: Iterative Refinement with Self-Feedback.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Self-Refine: Iterative Refinement with Self-Feedback

Reference 16

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.237462Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b41f381c-8c5c-41d9-8aef-74bdff93efee · outbound

This paper cites Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Adding Error Bars to Evals: A Statistical Approach to Language Model Evaluations

Reference 17

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.303898Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3f3b467a-fdfb-4b76-a18f-a7ceeddc2b93 · outbound

This paper cites Fair on the Surface: Transaction-Ordering Bias and MEV in Mysticeti DAG-based BFT Protocol.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Fair on the Surface: Transaction-Ordering Bias and MEV in Mysticeti DAG-based BFT Protocol

Reference 18

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.364677Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e40ff60f-e0e0-4c29-9860-597da010867b · outbound

This paper cites s1: Simple test-time scaling.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B s1: Simple test-time scaling

Reference 19

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.429895Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c09d33ea-403f-4738-85b1-ac4435c38fb5 · outbound

This paper cites Simplicity Paradox: Debunking myths about prompting and datasets for LLM evaluation.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Simplicity Paradox: Debunking myths about prompting and datasets for LLM evaluation

Reference 20

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.468509Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 849ec5dc-5993-463b-aab9-21ca3305c913 · outbound

This paper cites Qwen2.5 Technical Report.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Qwen2.5 Technical Report

Reference 21

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.517313Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ff981305-90ed-49c5-8259-b0a78b552d41 · outbound

This paper cites The sequential edge: Inverse-entropy voting beats parallel self-consistency at matched compute.arXiv preprint arXiv:2511.02309, 2025.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B The sequential edge: Inverse-entropy voting beats parallel self-consistency at matched compute.arXiv preprint arXiv:2511.02309, 2025

Reference 22

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.567426Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2a636804-bd97-416c-afc8-45d29c695e58 · outbound

This paper cites Reflexion: Language Agents with Verbal Reinforcement Learning.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Reflexion: Language Agents with Verbal Reinforcement Learning

Reference 23

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.608216Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3d58206-ae11-42d9-85a2-f42d2964c6e5 · outbound

This paper cites Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters

Reference 24

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.644343Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3e18785f-cc04-4bf4-b8eb-5619454463e1 · outbound

This paper cites Inference scaling flaws: The limits of llm resampling with imperfect verifiers.arXiv preprint arXiv:2411.17501, 2024.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Inference scaling flaws: The limits of llm resampling with imperfect verifiers.arXiv preprint arXiv:2411.17501, 2024

Reference 25

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.700425Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 68fa443e-7841-459f-814a-55147701f9bf · outbound

This paper cites Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Single-Agent LLMs Outperform Multi-Agent Systems on Multi-Hop Reasoning Under Equal Thinking Token Budgets

Reference 26

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.768803Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9e3773e0-13b6-4392-96de-f30f6bf94230 · outbound

This paper cites Reasoning in Token Economies: Budget-Aware Evaluation of LLM Reasoning Strategies.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Reasoning in Token Economies: Budget-Aware Evaluation of LLM Reasoning Strategies

Reference 27

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.810695Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1ab401c6-35bf-435e-9d8b-717772884b58 · outbound

This paper cites Plan-and-solve prompting: Improving zero-shot chain-of-thought reasoning by large language models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Plan-and-solve prompting: Improving zero-shot chain-of-thought reasoning by large language models

Reference 28

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.887948Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bb48a105-620c-44d4-8ad4-a06fb29c3e07 · outbound

This paper cites Self-Consistency Improves Chain of Thought Reasoning in Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Self-Consistency Improves Chain of Thought Reasoning in Language Models

Reference 29

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.035096Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8701c473-0ea7-40c1-b40a-557549e12a40 · outbound

This paper cites Chain-of-Thought Prompting Elicits Reasoning in Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Chain-of-Thought Prompting Elicits Reasoning in Large Language Models

Reference 30

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.111637Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f502b14f-dbfa-4269-9b97-4981a9bb318b · outbound

This paper cites Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Inference Scaling Laws: An Empirical Analysis of Compute-Optimal Inference for Problem-Solving with Language Models

Reference 31

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.177346Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3d636e9-6db7-40a2-88d0-9c5568733379 · outbound

This paper cites Tree of Thoughts: Deliberate Problem Solving with Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Tree of Thoughts: Deliberate Problem Solving with Large Language Models

Reference 32

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.220557Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 19fe31c5-9bf0-46b5-b97d-5da5dc5a9d70 · outbound

This paper cites Incentivizing llms to self-verify their answers.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Incentivizing llms to self-verify their answers

Reference 33

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.285736Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 92ca9c9d-b4e6-48ea-8da8-f6c54236be20 · outbound

This paper cites Progressive-Hint Prompting Improves Reasoning in Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Progressive-Hint Prompting Improves Reasoning in Large Language Models

Reference 34

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.329887Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 558abeb0-e40f-4fde-bca7-0c93e7a6f5f2 · outbound

This paper cites Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Reference 35

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.409610Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9f9128a1-3ea9-48a0-85e1-d1c939428a48 · outbound

This paper cites Least-to-Most Prompting Enables Complex Reasoning in Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Least-to-Most Prompting Enables Complex Reasoning in Large Language Models

Reference 36

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:01.477082Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b1bea29f-2a0c-4cc4-880a-5c2a2e311368 · outbound

This paper cites Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models.

Sample More, Reflect Less: Self-Refine and Reflexion Lose to Repeated Sampling at Equal Token Cost, from 1.5B to 7B Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Models

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-07-31T03:19:00.973897Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

No inbound Pith citation observations are available.