Paper Citation Record · LEDGER

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification

As of 10 August 2026, this Paper Citation Record lists 38 of 38 outbound references and 4 inbound Pith citation observations for arXiv:2502.06852.

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

pith.paper-citation-record.v1
2502.06852 v1

Coverage vector

measured 38 of 38 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-08T20:36:01.377075Z

measured 42 of 42 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-09T06:31:02.800959+00:00

measured 4 of 4 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-07T12:03:35.989818Z

measured 0 of 1 external citation measurements

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

Source: pith, observed 2026-08-06T23:59:54.421222Z

Reference resolution

38 of 38 outbound references displayed

verified exact2
verified fuzzy15
unresolved21
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation e8253413-3eb8-434a-8b44-ee9a1d913498 · outbound

This paper cites write newline.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification write newline

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.184239Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a9ea5a7e-60a7-4a8f-b8d8-d14604223173 · outbound

This paper cites GPT-4 Technical Report.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification GPT-4 Technical Report

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.192137Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cbb9d700-8b22-44ba-bd03-d5994e79132b · outbound

This paper cites Mechanistic Interpretability for AI Safety -- A Review.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Mechanistic Interpretability for AI Safety -- A Review

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.197844Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f44d8dce-fcd7-4b6a-b737-50ec65287529 · outbound

This paper cites Leveraging Logical Rules in Knowledge Editing: A Cherry on the Top.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Leveraging Logical Rules in Knowledge Editing: A Cherry on the Top

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.203022Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9fc47a8-dcb0-46fc-b006-3f14ce495833 · outbound

This paper cites Identifying and adapting transformer-components responsible for gender bias in an english language model.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Identifying and adapting transformer-components responsible for gender bias in an english language model

Reference 5

Resolution: verified exact; doi, observed 2026-08-08T20:36:01.434328Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 1c019116-47c9-44a9-acd9-6345b616ffd8 · outbound

This paper cites Towards automated circuit discovery for mechanistic interpretability.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Towards automated circuit discovery for mechanistic interpretability

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.162866Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 73eca2a6-492f-41f9-8e98-eeef50f76743 · outbound

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

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Sparse Autoencoders Find Highly Interpretable Features in Language Models

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.218839Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 4fb543e8-aa5d-41e9-b5a5-a1f0b69337d9 · outbound

This paper cites Bert: Pre-training of deep bidirectional transformers for language understanding.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Bert: Pre-training of deep bidirectional transformers for language understanding

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.144037Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation fb206b5c-2316-4bee-bf3b-d0755bcd9f29 · outbound

This paper cites Causal abstractions of neural networks.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Causal abstractions of neural networks

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.127495Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 83dd6122-a21a-4696-9dbd-9a10283705c1 · outbound

This paper cites Finding alignments between interpretable causal variables and distributed neural representations.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Finding alignments between interpretable causal variables and distributed neural representations

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.111229Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 883e4f08-97cb-411c-9ba3-fb900f292219 · outbound

This paper cites Transformer Feed-Forward Layers Are Key-Value Memories.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Transformer Feed-Forward Layers Are Key-Value Memories

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.239131Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c230c284-989f-43a7-93e0-90ceecdce87f · outbound

This paper cites How does gpt-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification How does gpt-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.087591Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation a5406c56-6867-4a48-a36c-7fb33b9bd850 · outbound

This paper cites Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Have Faith in Faithfulness: Going Beyond Circuit Overlap When Finding Model Mechanisms

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.249028Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d9fef52a-aaf3-4005-92ca-f840a473a382 · outbound

This paper cites Dictionary Learning Improves Patch-Free Circuit Discovery in Mechanistic Interpretability: A Case Study on Othello-GPT.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Dictionary Learning Improves Patch-Free Circuit Discovery in Mechanistic Interpretability: A Case Study on Othello-GPT

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.253940Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 81b8f35b-e56f-45c3-bad4-f981f9be4de4 · outbound

This paper cites Dissecting Fine-Tuning Unlearning in Large Language Models.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Dissecting Fine-Tuning Unlearning in Large Language Models

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.258787Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c0900e0c-3d82-4edc-b459-975a295a744d · outbound

This paper cites A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.263580Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b91ed11f-5e52-40a5-9839-c430ecb9e219 · outbound

This paper cites Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchilla.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchilla

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.268872Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 53dd1163-a2f3-4023-8604-395a5a90ed19 · outbound

This paper cites Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.274129Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e0460d97-cc06-474a-b014-56bc87091bb0 · outbound

This paper cites Locating and editing factual associations in gpt.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Locating and editing factual associations in gpt

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.067707Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 32763730-eec4-455d-a568-8cbbc48048d7 · outbound

This paper cites Circuit component reuse across tasks in transformer language models.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Circuit component reuse across tasks in transformer language models

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.044199Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 780707b0-73cd-47a6-9f9a-1dd4ebaa3ea0 · outbound

This paper cites Transformer Circuit Faithfulness Metrics are not Robust.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Transformer Circuit Faithfulness Metrics are not Robust

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.289787Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ac590c3b-4633-44b5-8e14-c3c9a630ac15 · outbound

This paper cites Evaluating Brain-Inspired Modular Training in Automated Circuit Discovery for Mechanistic Interpretability.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Evaluating Brain-Inspired Modular Training in Automated Circuit Discovery for Mechanistic Interpretability

Reference 22

Resolution: verified exact; local_arxiv, observed 2026-08-08T20:36:01.574229Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 2355de48-f761-4633-9328-935bce94252e · outbound

This paper cites Mechanistic interpretability quickstart guide.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Mechanistic interpretability quickstart guide

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:02.022969Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 87e6a542-f7e3-4cce-801c-1ce8a79b0deb · outbound

This paper cites Mechanistic interpretability, variables, and the importance of interpretable bases.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Mechanistic interpretability, variables, and the importance of interpretable bases

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.998500Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 98f741ce-392f-4c52-8a73-883809f4412a · outbound

This paper cites Zoom in: An introduction to circuits.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Zoom in: An introduction to circuits

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.971292Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 2dc6509b-abe5-4d5b-8f70-58aaef97af85 · outbound

This paper cites Sparse Autoencoders Enable Scalable and Reliable Circuit Identification in Language Models.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Sparse Autoencoders Enable Scalable and Reliable Circuit Identification in Language Models

Reference 26

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.317245Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0fc41602-6be4-48a6-aff1-0576a43855a8 · outbound

This paper cites Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Fine-Tuning Enhances Existing Mechanisms: A Case Study on Entity Tracking

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.323353Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 12d099cf-5a4d-4730-a891-3515cc03c21e · outbound

This paper cites Visualizing the impact of feature attribution baselines.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Visualizing the impact of feature attribution baselines

Reference 28

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.328749Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cc1e841e-ebb3-409d-92b2-bf33701673c1 · outbound

This paper cites Axiomatic attribution for deep networks.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Axiomatic attribution for deep networks

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.948151Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation db57e891-cce4-47b0-a567-c8cfa23c9035 · outbound

This paper cites Attribution Patching Outperforms Automated Circuit Discovery.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Attribution Patching Outperforms Automated Circuit Discovery

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.338244Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2b21d3cd-c995-4a02-9c98-9789c820c5d4 · outbound

This paper cites an unresolved cited work.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Unresolved cited work

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.343163Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2617758a-4828-40e9-8ad1-195bf894d2f9 · outbound

This paper cites Linear Representations of Sentiment in Large Language Models.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Linear Representations of Sentiment in Large Language Models

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.347727Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 75571146-ddba-4d3b-9c76-a249ad1efa30 · outbound

This paper cites N., Kaiser, L., and Polosukhin, I.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification N., Kaiser, L., and Polosukhin, I

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.912278Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 935bb76d-5614-4ba9-90fc-3732d3c2e878 · outbound

This paper cites Investigating gender bias in language models using causal mediation analysis.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Investigating gender bias in language models using causal mediation analysis

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.893975Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 68fab141-3734-42b7-ab71-a33fd692fafa · outbound

This paper cites R., Variengien, A., Conmy, A., Shlegeris, B., and Steinhardt, J.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification R., Variengien, A., Conmy, A., Shlegeris, B., and Steinhardt, J

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.870968Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 79e53d76-d252-49e7-bcf3-168f72281644 · outbound

This paper cites Interpretability at scale: Identifying causal mechanisms in alpaca.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Interpretability at scale: Identifying causal mechanisms in alpaca

Reference 36

Resolution: verified fuzzy; raw_fallback, observed 2026-08-08T20:36:01.854179Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation 23f1268e-f1df-4f4c-b6ad-fb45ea0e3bf2 · outbound

This paper cites Exploring the Personality Traits of LLMs through Latent Features Steering.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Exploring the Personality Traits of LLMs through Latent Features Steering

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.371974Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ef825e48-ef89-44d2-97be-742c75da76c2 · outbound

This paper cites Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing.

EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification Locate-then-edit for Multi-hop Factual Recall under Knowledge Editing

Reference 38

Resolution: unresolved; no resolver link, observed 2026-08-08T20:36:01.377075Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Pith citing papers

Observation 858c688b-1ce2-432c-9ddc-508c64e0c5ce · inbound

Understanding and Mitigating Cross-lingual Privacy Leakage via Language-specific and Universal Privacy Neurons cites this paper.

Understanding and Mitigating Cross-lingual Privacy Leakage via Language-specific and Universal Privacy Neurons EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification

Reference 55

Resolution: unresolved; no resolver link, observed 2026-08-07T12:03:35.989818Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e3eae1d4-6458-49af-a210-25a370eed43b · inbound

COMPKE: Complex Question Answering under Knowledge Editing cites this paper.

COMPKE: Complex Question Answering under Knowledge Editing EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification

Reference 60

Resolution: unresolved; no resolver link, observed 2026-08-07T12:00:23.857265Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 71a0bbe3-051d-469c-8678-10346f8e9a81 · inbound

The Compositional Architecture of Regret in Large Language Models cites this paper.

The Compositional Architecture of Regret in Large Language Models EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification

Reference 8

Resolution: verified exact; local_arxiv, observed 2026-08-06T23:59:54.489369Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-09T06:31:02.800959+00:00.

Observation f3c46278-2027-4a37-9d16-ccf65a3ffe56 · inbound

Query Circuits: Explaining How Language Models Answer User Prompts cites this paper.

Query Circuits: Explaining How Language Models Answer User Prompts EAP-GP: Mitigating Saturation Effect in Gradient-based Automated Circuit Identification

Reference 51

Resolution: unresolved; no resolver link, observed 2026-08-04T13:54:29.774749Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.