Paper Citation Record · LEDGER

Linear Probe Penalties Reduce LLM Sycophancy

As of 23 August 2026, this Paper Citation Record lists 34 of 34 outbound references and 7 inbound Pith citation observations for arXiv:2412.00967.

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

pith.paper-citation-record.v1
2412.00967 v1

Coverage vector

measured 34 of 34 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-12T04:53:06.503806Z

measured 41 of 41 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-23T06:30:58.430688+00:00

measured 7 of 7 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-08T05:19:12.781624Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-07-03T00:47:29.997154Z

Reference resolution

34 of 34 outbound references displayed

verified exact0
verified fuzzy11
unresolved23
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 8b710cbf-ab86-4b3f-bcfb-c14b113e4887 · outbound

This paper cites Understanding intermediate layers using linear classifier probes.

Linear Probe Penalties Reduce LLM Sycophancy Understanding intermediate layers using linear classifier probes

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.372154Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 38caefae-1b8a-487c-b65e-8e9ae02132f5 · outbound

This paper cites A positivity bias in written and spoken english and its moderation by personality and gender.

Linear Probe Penalties Reduce LLM Sycophancy A positivity bias in written and spoken english and its moderation by personality and gender

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.890407Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 760992be-f392-4d6b-a37a-56f54cb1acc4 · outbound

This paper cites Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback.

Linear Probe Penalties Reduce LLM Sycophancy Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.380667Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c036c875-d4e1-40e7-b017-e0cde7eb336d · outbound

This paper cites On the Opportunities and Risks of Foundation Models.

Linear Probe Penalties Reduce LLM Sycophancy On the Opportunities and Risks of Foundation Models

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.384899Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b38d39b0-c898-4b59-9a53-8d25da4c4240 · outbound

This paper cites Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback.

Linear Probe Penalties Reduce LLM Sycophancy Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.389428Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b30d288e-1db0-4b2d-92f3-eb718c909ab5 · outbound

This paper cites Deep reinforcement learning from human preferences.

Linear Probe Penalties Reduce LLM Sycophancy Deep reinforcement learning from human preferences

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.393513Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a19cbed6-4c23-4db1-a75c-5969f5f28716 · outbound

This paper cites UltraFeedback: Boosting language models with high-quality feedback,.

Linear Probe Penalties Reduce LLM Sycophancy UltraFeedback: Boosting language models with high-quality feedback,

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.870440Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 592b7e19-a67b-4d82-a0a3-ab34ed448b11 · outbound

This paper cites Human language reveals a universal positivity bias.

Linear Probe Penalties Reduce LLM Sycophancy Human language reveals a universal positivity bias

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.858415Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation d6e4e40f-f362-4a43-b8bf-052f85703564 · outbound

This paper cites an unresolved cited work.

Linear Probe Penalties Reduce LLM Sycophancy Unresolved cited work

Reference 9

Resolution: unresolved; raw_fallback, observed 2026-08-12T04:53:06.847453Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation f8a815df-0e3e-4948-ab7e-b3ee4c8dc1c7 · outbound

This paper cites siebert/sentiment-roberta-large-english · hugging face, 2021.

Linear Probe Penalties Reduce LLM Sycophancy siebert/sentiment-roberta-large-english · hugging face, 2021

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.836017Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 30c074fc-d465-468d-a45e-390d2eb8d031 · outbound

This paper cites LLM Agents can Autonomously Hack Websites.

Linear Probe Penalties Reduce LLM Sycophancy LLM Agents can Autonomously Hack Websites

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.417443Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4653fe7c-b576-44d6-9454-969a87b5df0c · outbound

This paper cites Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned.

Linear Probe Penalties Reduce LLM Sycophancy Red Teaming Language Models to Reduce Harms: Methods, Scaling Behaviors, and Lessons Learned

Reference 12

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.421716Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fa25cec5-b362-4418-acba-4aa39a577d84 · outbound

This paper cites Interference of the end: Why recency bias in memory determines when a food is consumed again.

Linear Probe Penalties Reduce LLM Sycophancy Interference of the end: Why recency bias in memory determines when a food is consumed again

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.824191Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation dfc429e2-4a5b-4a68-bef5-59f788c8e62a · outbound

This paper cites More than a feeling: Accuracy and application of sentiment analysis.

Linear Probe Penalties Reduce LLM Sycophancy More than a feeling: Accuracy and application of sentiment analysis

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.428918Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 75072bf4-7d34-46c6-bbcd-23118cbf1548 · outbound

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

Linear Probe Penalties Reduce LLM Sycophancy Large Language Models are Zero-Shot Reasoners

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.433407Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 27c2431a-ea83-4f41-b054-08c60154f045 · outbound

This paper cites Still No Lie Detector for Language Models: Probing Empirical and Conceptual Roadblocks.

Linear Probe Penalties Reduce LLM Sycophancy Still No Lie Detector for Language Models: Probing Empirical and Conceptual Roadblocks

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.437344Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cd14188e-5db7-4734-b332-363ad7596ad8 · outbound

This paper cites Emergent linear representations in world models of self-supervised sequence models, 2023.

Linear Probe Penalties Reduce LLM Sycophancy Emergent linear representations in world models of self-supervised sequence models, 2023

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.813024Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation c0449222-c72d-426f-8eb3-21cfed7a75e3 · outbound

This paper cites Training language models to follow instructions with human feedback.

Linear Probe Penalties Reduce LLM Sycophancy Training language models to follow instructions with human feedback

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.444875Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 73e62d59-20a2-4170-a8cf-3153355c4d0b · outbound

This paper cites Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales.

Linear Probe Penalties Reduce LLM Sycophancy Seeing stars: Exploiting class relationships for sentiment categorization with respect to rating scales

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.448384Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b2bec276-2a96-4c61-a557-1c69ab9b8471 · outbound

This paper cites Prioritizing High-Consequence Biological Capabilities in Evaluations of Artificial Intelligence Models.

Linear Probe Penalties Reduce LLM Sycophancy Prioritizing High-Consequence Biological Capabilities in Evaluations of Artificial Intelligence Models

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.452479Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6d66ed55-9aec-4604-b8d9-da3a17c68fa6 · outbound

This paper cites Discovering Language Model Behaviors with Model-Written Evaluations.

Linear Probe Penalties Reduce LLM Sycophancy Discovering Language Model Behaviors with Model-Written Evaluations

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.456194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 58a24925-d63e-4101-8aa4-2293805d0540 · outbound

This paper cites Steering Llama 2 via Contrastive Activation Addition.

Linear Probe Penalties Reduce LLM Sycophancy Steering Llama 2 via Contrastive Activation Addition

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.459992Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 22b701c6-153e-4978-b58a-63b786c416cb · outbound

This paper cites Bowman, Newton Cheng, Esin Durmus, Zac Hatfield-Dodds, Scott R.

Linear Probe Penalties Reduce LLM Sycophancy Bowman, Newton Cheng, Esin Durmus, Zac Hatfield-Dodds, Scott R

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.794428Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation c4051584-aa8f-4bdd-856a-678f4f1f69ca · outbound

This paper cites Immediate and delayed primacy and recency effects in performance evaluation.

Linear Probe Penalties Reduce LLM Sycophancy Immediate and delayed primacy and recency effects in performance evaluation

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.782841Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation b7a91471-2f7f-4fb4-9e85-fa3fb203032b · outbound

This paper cites Towards Understanding Sycophancy in Language Models.

Linear Probe Penalties Reduce LLM Sycophancy Towards Understanding Sycophancy in Language Models

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.468357Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 39b824f1-c9d9-4d28-8a48-a1739e65bf59 · outbound

This paper cites Simple synthetic data reduces sycophancy in large language models.

Linear Probe Penalties Reduce LLM Sycophancy Simple synthetic data reduces sycophancy in large language models

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.479722Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 31dc03a1-f3fd-411f-87d3-e659b81c1d3d · outbound

This paper cites OpenChat: Advancing Open-source Language Models with Mixed-Quality Data.

Linear Probe Penalties Reduce LLM Sycophancy OpenChat: Advancing Open-source Language Models with Mixed-Quality Data

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.475789Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f3549866-c168-4a52-9ae9-47e7247b269b · outbound

This paper cites as a judge.

Linear Probe Penalties Reduce LLM Sycophancy as a judge

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.762121Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation cfdd1245-6589-4265-8565-8991d0fb87cb · outbound

This paper cites Starling-7b: Improving llm helpfulness & harmlessness with rlaif, 2023.

Linear Probe Penalties Reduce LLM Sycophancy Starling-7b: Improving llm helpfulness & harmlessness with rlaif, 2023

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.483661Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cda8f650-92d4-4d7a-b2df-01d42b317860 · outbound

This paper cites an unresolved cited work.

Linear Probe Penalties Reduce LLM Sycophancy Unresolved cited work

Reference 31

Resolution: unresolved; raw_fallback, observed 2026-08-12T04:53:06.750568Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 9ecb7f78-a3bf-454b-9a14-2dda37db3fa4 · outbound

This paper cites (B) Positive.

Linear Probe Penalties Reduce LLM Sycophancy (B) Positive

Reference 32

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.738779Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation dfc90fe1-cd57-4c9f-b170-393e68afd637 · outbound

This paper cites an unresolved cited work.

Linear Probe Penalties Reduce LLM Sycophancy Unresolved cited work

Reference 33

Resolution: unresolved; raw_fallback, observed 2026-08-12T04:53:06.727473Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 682f048f-637d-4093-9789-ae5e8c50fb52 · outbound

This paper cites The distribution is now much more balanced.

Linear Probe Penalties Reduce LLM Sycophancy The distribution is now much more balanced

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-12T04:53:06.715215Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 3d8a41a7-74b7-47c1-85ae-02c0e8f156ef · outbound

This paper cites UltraFeedback: Boosting Language Models with Scaled AI Feedback.

Linear Probe Penalties Reduce LLM Sycophancy UltraFeedback: Boosting Language Models with Scaled AI Feedback

Reference 2023

Resolution: unresolved; no resolver link, observed 2026-08-12T04:53:06.401822Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 36c36c31-3b73-457e-92c9-4919fbb3cb87 · inbound

LPASS: Linear Probes as Stepping Stones for vulnerability detection using compressed LLMs cites this paper.

LPASS: Linear Probes as Stepping Stones for vulnerability detection using compressed LLMs Linear Probe Penalties Reduce LLM Sycophancy

Reference 42

Resolution: unresolved; no resolver link, observed 2026-08-07T12:28:30.431171Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4e3323b6-488b-486c-a837-1b926db7d428 · inbound

Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework cites this paper.

Manipulation Attacks by Misaligned AI: Risk Analysis and Safety Case Framework Linear Probe Penalties Reduce LLM Sycophancy

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-06T16:39:36.959394Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2441dda6-046a-401f-a95b-2d7c6ac10835 · inbound

CausalT5k: Diagnosing Refusal and Failure Modes in Trustworthy Causal Reasoning Across Causal Rungs cites this paper.

CausalT5k: Diagnosing Refusal and Failure Modes in Trustworthy Causal Reasoning Across Causal Rungs Linear Probe Penalties Reduce LLM Sycophancy

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-03T03:09:49.308697Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3da1a77b-b09a-4827-a2df-25dcd9cec6b9 · inbound

Pressure, What Pressure? Sycophancy Disentanglement in Language Models via Reward Decomposition cites this paper.

Pressure, What Pressure? Sycophancy Disentanglement in Language Models via Reward Decomposition Linear Probe Penalties Reduce LLM Sycophancy

Reference 6

Resolution: verified exact; arxiv_id, observed 2026-05-10T22:10:49.183370Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 56a958e8-e3e7-4eb0-964a-754c28de3fec · inbound

Beyond Semantic Relevance: Counterfactual Risk Minimization for Robust Retrieval-Augmented Generation cites this paper.

Beyond Semantic Relevance: Counterfactual Risk Minimization for Robust Retrieval-Augmented Generation Linear Probe Penalties Reduce LLM Sycophancy

Reference 62

Resolution: verified exact; arxiv_id, observed 2026-05-11T16:46:11.336744Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation 31dbaf3a-b5c2-4511-b779-91f71efdcd61 · inbound

Emergent Misalignment Can Be Induced by Sycophancy and Reversed via Alignment Gating cites this paper.

Emergent Misalignment Can Be Induced by Sycophancy and Reversed via Alignment Gating Linear Probe Penalties Reduce LLM Sycophancy

Reference 21

Resolution: metadata mismatch; arxiv_id, observed 2026-07-03T00:47:29.998699Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-23T06:30:58.430688+00:00.

Observation f5594707-a557-43eb-a7e3-9b2adc876ff0 · inbound

Measuring and Detecting Harmful AI Sycophancy cites this paper.

Measuring and Detecting Harmful AI Sycophancy Linear Probe Penalties Reduce LLM Sycophancy

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-08T05:19:12.781624Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.