Paper Citation Record · LEDGER

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training

As of 7 August 2026, this Paper Citation Record lists 62 of 62 outbound references and 1 inbound Pith citation observation for arXiv:2507.12759.

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

pith.paper-citation-record.v1
2507.12759 v1

Coverage vector

measured 62 of 62 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-06T16:46:01.590792Z

measured 63 of 63 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-06T06:34:29.942622+00:00

measured 1 of 1 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-04T09:12:10.059856Z

measured 0 of 1 external citation measurements

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

Source: cited_works

Reference resolution

62 of 62 outbound references displayed

verified exact3
verified fuzzy13
unresolved45
parse uncertain0
malformed identifier1
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 21997aab-3d1a-4cb8-b179-e2f83a94a126 · outbound

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

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Large Language Monkeys: Scaling Inference Compute with Repeated Sampling

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.445386Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f6c26582-69dd-4d78-b5a2-713aefe9aad1 · outbound

This paper cites Evaluating Large Language Models Trained on Code.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Evaluating Large Language Models Trained on Code

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.493664Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 12c0f1fe-df2f-42d6-a4e0-8ffaec5705d1 · outbound

This paper cites Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.582452Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation de9cfe34-37b3-43fc-ab64-454abe207e69 · outbound

This paper cites Glass, and Pengcheng He.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Glass, and Pengcheng He

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:04.279684Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation a9939c19-8436-4e7b-b368-9e44db16b45e · outbound

This paper cites Training Verifiers to Solve Math Word Problems.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Training Verifiers to Solve Math Word Problems

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.752932Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2c64b3b2-1519-44f8-8d23-cfdb3e01f7ff · outbound

This paper cites Reinforcement learning for reasoning in small llms: What works and what doesn't.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Reinforcement learning for reasoning in small llms: What works and what doesn't

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.817307Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 683c4a71-7ef6-48e9-b07e-7feec6730d36 · outbound

This paper cites DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:56.908594Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8f246712-f3af-47f5-9297-06a5801a5a63 · outbound

This paper cites On giant's shoulders: Effortless weak to strong by dynamic logits fusion.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training On giant's shoulders: Effortless weak to strong by dynamic logits fusion

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:04.110157Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 72107da5-6899-40ae-b762-4edcbbd86c67 · outbound

This paper cites On giant's shoulders: Effortless weak to strong by dynamic logits fusion.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training On giant's shoulders: Effortless weak to strong by dynamic logits fusion

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:04.001284Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 70215125-4178-4489-98dd-6b7077763c63 · outbound

This paper cites Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.160987Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9b39ac8c-3f01-4e6e-a3fb-269e31ca2629 · outbound

This paper cites Measuring mathematical problem solving with the MATH dataset.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Measuring mathematical problem solving with the MATH dataset

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.244757Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f2f772cb-6146-4cc0-9396-c32185c08b19 · outbound

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

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training V-STaR: Training Verifiers for Self-Taught Reasoners

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.308861Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 02eb302f-a275-47bb-a55e-eeeb33530590 · outbound

This paper cites Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen - Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen - Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.397518Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0573bae5-1f12-4cf5-bfb5-eb62573307aa · outbound

This paper cites Huang, Wenxuan Zhou, Fei Wang, Fred Morstatter, Sheng Zhang, Hoifung Poon, and Muhao Chen.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Huang, Wenxuan Zhou, Fei Wang, Fred Morstatter, Sheng Zhang, Hoifung Poon, and Muhao Chen

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.887735Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 6df1b769-c361-48cb-89dc-a0ddd3936064 · outbound

This paper cites GRACE: discriminator-guided chain-of-thought reasoning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training GRACE: discriminator-guided chain-of-thought reasoning

Reference 15

Resolution: verified exact; doi, observed 2026-08-06T16:46:02.242062Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 3621be0a-5760-4cd0-acf6-9e790bf65ca4 · outbound

This paper cites Process reward models that think.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Process reward models that think

Reference 16

Resolution: verified exact; doi, observed 2026-08-06T16:46:02.097358Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 3cf58925-69cf-438d-894e-8ad8100b90dc · outbound

This paper cites Tulu 3: Pushing Frontiers in Open Language Model Post-Training.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Tulu 3: Pushing Frontiers in Open Language Model Post-Training

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.751295Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 30810393-db81-491d-b1e5-367ac46a5321 · outbound

This paper cites Fast inference from transformers via speculative decoding.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Fast inference from transformers via speculative decoding

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.730256Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation a2c1abac-b546-4a5d-91d8-a572085d039e · outbound

This paper cites LLMs Can Easily Learn to Reason from Demonstrations Structure, not content, is what matters!.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training LLMs Can Easily Learn to Reason from Demonstrations Structure, not content, is what matters!

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:57.928076Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9b47a880-1dc7-4fec-a121-dbaa9425df30 · outbound

This paper cites Contrastive decoding: Open-ended text generation as optimization.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Contrastive decoding: Open-ended text generation as optimization

Reference 20

Resolution: malformed identifier; no resolver link, observed 2026-08-06T16:45:58.052123Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c7e65240-89f6-420b-8825-cc760c589dcf · outbound

This paper cites Reward-Guided Speculative Decoding for Efficient LLM Reasoning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Reward-Guided Speculative Decoding for Efficient LLM Reasoning

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.173827Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b3105c22-e2b1-4ff2-92fc-b78830b46e74 · outbound

This paper cites Let's verify step by step.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Let's verify step by step

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.268945Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f5267fe8-49f1-4b47-b047-d02ad14612b3 · outbound

This paper cites Smith, and Yejin Choi.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Smith, and Yejin Choi

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.360990Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 48185574-370d-49dd-8c9f-6ea8dab5867c · outbound

This paper cites Tuning Language Models by Proxy.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Tuning Language Models by Proxy

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.459933Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d1c89f52-4375-4116-8259-c195484abd60 · outbound

This paper cites Understanding R1-Zero-Like Training: A Critical Perspective.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Understanding R1-Zero-Like Training: A Critical Perspective

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.575606Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation de81d482-37a6-48e5-a4fc-1cd76d878270 · outbound

This paper cites Tang, Manan Roongta, Colin Cai, Jeffrey Luo, Li Erran Li, Raluca Ada Popa, and Ion Stoica.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Tang, Manan Roongta, Colin Cai, Jeffrey Luo, Li Erran Li, Raluca Ada Popa, and Ion Stoica

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.615685Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 6cfea145-9d9b-4e0e-807a-7a4f0c3969ac · outbound

This paper cites an unresolved cited work.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Unresolved cited work

Reference 27

Resolution: unresolved; raw_fallback, observed 2026-08-06T16:46:03.498018Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 4e47f8da-12a8-42af-b74d-bf6a60355d36 · outbound

This paper cites s1: Simple test-time scaling.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training s1: Simple test-time scaling

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:58.836736Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 08f789f8-0a12-4fc8-b284-41784ef7ce7c · outbound

This paper cites Learning to reason with LLMs , September 2024.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Learning to reason with LLMs , September 2024

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.388308Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation a6b54b39-9210-4384-ac66-08f1b3aa6d00 · outbound

This paper cites OpenAI o3 and o4-mini System Card.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training OpenAI o3 and o4-mini System Card

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.272016Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 4c76ae28-7628-4c96-a283-74aff655653e · outbound

This paper cites Comblm: Adapting black-box language models through small fine-tuned models.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Comblm: Adapting black-box language models through small fine-tuned models

Reference 31

Resolution: verified exact; doi, observed 2026-08-06T16:46:01.825336Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 840be525-fc0a-4339-8be2-f5cfaa0a7ec1 · outbound

This paper cites Qwen3: Think Deeper, Act Faster | Qwen , April 2025.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Qwen3: Think Deeper, Act Faster | Qwen , April 2025

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:03.125136Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 310ce96a-aba3-4598-a222-1e967718ed69 · outbound

This paper cites Manning, Stefano Ermon, and Chelsea Finn.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Manning, Stefano Ermon, and Chelsea Finn

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.275958Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3b16074f-bd1c-418b-9eb6-a96d2e40d9a3 · outbound

This paper cites DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

Reference 34

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.339469Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f117e112-6122-43f6-a1cb-58c24846e449 · outbound

This paper cites Trusting your evidence: Hallucinate less with context-aware decoding.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Trusting your evidence: Hallucinate less with context-aware decoding

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.448540Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 86ee16f0-2eb0-4833-9c2c-61f33e7445a3 · outbound

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

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.557416Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52fc4979-7cf7-4065-99ee-5faba847a214 · outbound

This paper cites Unlocking General Long Chain-of-Thought Reasoning Capabilities of Large Language Models via Representation Engineering.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Unlocking General Long Chain-of-Thought Reasoning Capabilities of Large Language Models via Representation Engineering

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.645017Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dab5bc7a-86df-4001-a0f0-1f33cda3ed40 · outbound

This paper cites Qwq-32b: Embracing the power of reinforcement learning, March 2025.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Qwq-32b: Embracing the power of reinforcement learning, March 2025

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.731670Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a204399-44e7-4aee-bac5-befa0570377e · outbound

This paper cites Gram: A generative foundation reward model for reward generalization.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Gram: A generative foundation reward model for reward generalization

Reference 39

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.800747Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4d807b13-d1dd-4180-8de4-03d264afaa1a · outbound

This paper cites Math-shepherd: Verify and reinforce llms step-by-step without human annotations.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Math-shepherd: Verify and reinforce llms step-by-step without human annotations

Reference 40

Resolution: unresolved; no resolver link, observed 2026-08-06T16:45:59.879033Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0a44f2d9-2071-4928-8e47-1f899fa880d6 · outbound

This paper cites Chain-of-thought reasoning without prompting.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Chain-of-thought reasoning without prompting

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:02.991227Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 5feaae47-f273-4fef-a58b-f96b94a4dbb5 · outbound

This paper cites Le, Ed H.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Le, Ed H

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:02.845097Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation 2516df83-9dad-4a49-b11f-032b8b455817 · outbound

This paper cites Reinforcement Learning for Reasoning in Large Language Models with One Training Example.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Reinforcement Learning for Reasoning in Large Language Models with One Training Example

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.145099Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2528988e-4912-4c24-8b27-f7042a856eb2 · outbound

This paper cites Chi, Quoc V.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Chi, Quoc V

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.198876Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7e621c17-b2d2-44fd-866a-9e7f7dd2ef07 · outbound

This paper cites Self-evaluation guided beam search for reasoning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Self-evaluation guided beam search for reasoning

Reference 45

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:02.724881Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation b8cd5682-e691-4445-a0c9-5039ab6d027c · outbound

This paper cites RedStar: Does Scaling Long-CoT Data Unlock Better Slow-Reasoning Systems?.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training RedStar: Does Scaling Long-CoT Data Unlock Better Slow-Reasoning Systems?

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.345444Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1a7c90e7-3566-4dcf-b080-56aefd0a6be4 · outbound

This paper cites Qwen2.5 Technical Report.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Qwen2.5 Technical Report

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.453742Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5f69db3d-332f-4eee-8f94-050a76d6dc61 · outbound

This paper cites Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement

Reference 48

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.536206Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7d94f4e8-e7dc-4d79-a4e4-9fb7ed4b21c4 · outbound

This paper cites Speculative Thinking: Enhancing Small-Model Reasoning with Large Model Guidance at Inference Time.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Speculative Thinking: Enhancing Small-Model Reasoning with Large Model Guidance at Inference Time

Reference 49

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.618572Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1d121914-4e0e-4a78-b5b5-dce25e4aa682 · outbound

This paper cites LIMO: Less is More for Reasoning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training LIMO: Less is More for Reasoning

Reference 50

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.686246Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b8b882c6-825e-46c9-b0d5-87a17175121c · outbound

This paper cites DAPO: An Open-Source LLM Reinforcement Learning System at Scale.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training DAPO: An Open-Source LLM Reinforcement Learning System at Scale

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.730849Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dc8ecd4c-f1e4-4f32-823c-db3c9e3cf9c6 · outbound

This paper cites Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?

Reference 52

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.822409Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4b1d8fbd-71c1-478d-9eeb-d4041f9c5445 · outbound

This paper cites Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:00.912514Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4f63ac81-53f5-45a3-b5dd-0df1df80a4c0 · outbound

This paper cites Generative verifiers: Reward modeling as next-token prediction.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Generative verifiers: Reward modeling as next-token prediction

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-08-06T16:46:02.603535Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-06T06:34:29.942622+00:00.

Observation db22c645-3a9d-4449-ae2f-e80f46d7454a · outbound

This paper cites Small language models need strong verifiers to self-correct reasoning.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Small language models need strong verifiers to self-correct reasoning

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.063828Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2687ee65-1b68-46c1-91c9-905b049d5d40 · outbound

This paper cites Weak-to-Strong Jailbreaking on Large Language Models.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Weak-to-Strong Jailbreaking on Large Language Models

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.155732Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 35d16542-ddcf-4da5-ab10-74fa647dbe5e · outbound

This paper cites Activation Control for Efficiently Eliciting Long Chain-of-thought Ability of Language Models.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Activation Control for Efficiently Eliciting Long Chain-of-thought Ability of Language Models

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.206824Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8a7f6206-36ff-4e1b-905e-4c83ef8e8ff2 · outbound

This paper cites Representation Engineering: A Top-Down Approach to AI Transparency.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Representation Engineering: A Top-Down Approach to AI Transparency

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.298001Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a1851c04-4c34-4986-ba23-3b96d3de57a1 · outbound

This paper cites write newline.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training write newline

Reference 59

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.378453Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dde975b2-07f5-4502-ab2f-bc6bc3920c34 · outbound

This paper cites @esa (Ref.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training @esa (Ref

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.466254Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae5ffced-7fb6-493b-ab02-3f51676a415a · outbound

This paper cites an unresolved cited work.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Unresolved cited work

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.530230Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c720fe2c-13ba-4858-bcc3-25f0e614f163 · outbound

This paper cites an unresolved cited work.

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training Unresolved cited work

Reference 62

Resolution: unresolved; no resolver link, observed 2026-08-06T16:46:01.590792Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 3832b309-32bc-404f-8095-26e1c51ab8a8 · inbound

Video Reasoning without Training cites this paper.

Video Reasoning without Training Logit Arithmetic Elicits Long Reasoning Capabilities Without Training

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-04T09:12:10.059856Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.