Paper Citation Record · LEDGER

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders

As of 17 August 2026, this Paper Citation Record lists 27 of 27 outbound references and 0 inbound Pith citation observations for arXiv:2608.08168.

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

pith.paper-citation-record.v1
2608.08168 v1

Coverage vector

measured 27 of 27 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-12T00:23:50.523918Z

measured 27 of 27 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-17T06:30:58.91139+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

27 of 27 outbound references displayed

verified exact0
verified fuzzy6
unresolved21
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 9cbcc900-a783-4fb0-9e27-92102fc38583 · outbound

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

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.418419Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 18ea0f4e-bc63-4f89-859c-f3711b44f138 · outbound

This paper cites Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.423415Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1b8f143f-e2be-4612-b50f-c6a9e03c2836 · outbound

This paper cites Language models don’t always say what they think: Unfaithful explanations in chain-of-thought prompting.Advances in Neural Information Processing Systems, 36:74952–74965, 2023.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Language models don’t always say what they think: Unfaithful explanations in chain-of-thought prompting.Advances in Neural Information Processing Systems, 36:74952–74965, 2023

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.427827Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 290a74bd-ce7a-4316-875e-ad22de85c5d3 · outbound

This paper cites Let’s verify step by step.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Let’s verify step by step

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.431998Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 33f4f543-12ee-479c-82d3-1435710c04de · outbound

This paper cites How does Chain of Thought Think? Mechanistic Interpretability of Chain-of-Thought Reasoning with Sparse Autoencoding.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders How does Chain of Thought Think? Mechanistic Interpretability of Chain-of-Thought Reasoning with Sparse Autoencoding

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.436202Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 25f08bb4-d134-458b-8514-0b0de5a37b2b · outbound

This paper cites Finding sparse autoencoder representations of errors in cot prompting.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Finding sparse autoencoder representations of errors in cot prompting

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.891768Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 2d4e165a-6dc8-4c43-90b0-a2c62a911c0e · outbound

This paper cites Progress measures for grokking via mechanistic interpretability.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Progress measures for grokking via mechanistic interpretability

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.445426Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3007f258-dc91-4983-88cd-871ce36b1c59 · outbound

This paper cites Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.449277Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 87bd8101-a35a-44b4-92c1-0f84e1adcae6 · outbound

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

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Representation Engineering: A Top-Down Approach to AI Transparency

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.453594Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5a1bb2e0-1002-4714-95ed-e3b8717a381a · outbound

This paper cites Improving Dictionary Learning with Gated Sparse Autoencoders.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Improving Dictionary Learning with Gated Sparse Autoencoders

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.457846Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8328cf95-7b15-438b-b4bb-a9189be04b13 · outbound

This paper cites Sparse Autoencoders Find Highly Interpretable Features in Language Models.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Sparse Autoencoders Find Highly Interpretable Features in Language Models

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.462347Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dc9a124f-eb76-40b0-83f5-827bb157a498 · outbound

This paper cites Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread, 2, 2023.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Towards monosemanticity: Decomposing language models with dictionary learning.Transformer Circuits Thread, 2, 2023

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.466491Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 115a9d7f-3f22-4362-9688-271f013aea15 · outbound

This paper cites Feature Extraction and Steering for Enhanced Chain-of-Thought Reasoning in Language Models.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Feature Extraction and Steering for Enhanced Chain-of-Thought Reasoning in Language Models

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.470682Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6b0e1e58-301d-4bf1-ae33-d7a88e444086 · outbound

This paper cites Locating and editing factual associations in gpt.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Locating and editing factual associations in gpt

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.475049Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 27b113c0-62ab-4821-9d15-3119b15046de · outbound

This paper cites Scaling and evaluating sparse autoencoders.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Scaling and evaluating sparse autoencoders

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.478491Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ada82d09-0663-421c-a985-c39ccd67f7e3 · outbound

This paper cites Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.482648Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 62dbf1e2-29f5-4e4e-8a34-db0cdfac515b · outbound

This paper cites Sparse autoencoder features for classifications and transferability.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Sparse autoencoder features for classifications and transferability

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.864087Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 27f7208c-d0eb-4fdf-806f-3f14ac1c762e · outbound

This paper cites Saes are good for steering–if you select the right features.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Saes are good for steering–if you select the right features

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.490578Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ece60cbd-713e-4dd9-a3dd-e999016a653b · outbound

This paper cites Lingualens: Towards interpreting linguistic mechanisms of large language models via sparse auto-encoder.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Lingualens: Towards interpreting linguistic mechanisms of large language models via sparse auto-encoder

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.851200Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation 14b9792e-6e85-4ffc-87fe-a920bfaaa251 · outbound

This paper cites Decoding Dense Embeddings: Sparse Autoencoders for Interpreting and Discretizing Dense Retrieval.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Decoding Dense Embeddings: Sparse Autoencoders for Interpreting and Discretizing Dense Retrieval

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.498053Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9a9da866-17ca-4060-9735-6573655149a4 · outbound

This paper cites k-Sparse Autoencoders.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders k-Sparse Autoencoders

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.501902Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 263e575a-3be9-4270-84bf-193f546a2271 · outbound

This paper cites DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.505227Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e503d8e4-528e-487f-9c43-5f1f235040d1 · outbound

This paper cites Reasoning Models Can Be Effective Without Thinking.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Reasoning Models Can Be Effective Without Thinking

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.508557Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 53ef2980-cb57-4d52-83ca-8bf913379abd · outbound

This paper cites Interpreting and steering llm representations with mutual information-based explanations on sparse autoencoders.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Interpreting and steering llm representations with mutual information-based explanations on sparse autoencoders

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.836653Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation b85240b3-6c73-4274-b747-c8c9a8e3cf3d · outbound

This paper cites Adam: A Method for Stochastic Optimization.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Adam: A Method for Stochastic Optimization

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-12T00:23:50.516505Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eb2de8b2-28f3-4360-a468-ce05979d74fb · outbound

This paper cites Scaling monosemanticity: Extracting interpretable features from claude 3 sonnet.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders Scaling monosemanticity: Extracting interpretable features from claude 3 sonnet

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.821013Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Observation a8016092-872c-4628-bf30-8c059e3c4502 · outbound

This paper cites wait", "hmm.

Thinking vs. NoThinking: Towards Interpreting Reasoning Mechanisms of Large Language Models via Sparse Autoencoders wait", "hmm

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T00:23:50.806064Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-17T06:30:58.91139+00:00.

Pith citing papers

No inbound Pith citation observations are available.