Paper Citation Record · LEDGER

Crafting Reversible SFT Behaviors in Large Language Models

As of 6 August 2026, this Paper Citation Record lists 40 of 40 outbound references and 0 inbound Pith citation observations for arXiv:2605.06632.

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

pith.paper-citation-record.v1
2605.06632 v1

Coverage vector

measured 40 of 40 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-05-08T12:16:16.104176Z

measured 40 of 40 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 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

40 of 40 outbound references displayed

verified exact22
verified fuzzy18
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation 212cdefd-ab25-4d43-a896-082f513fb478 · outbound

This paper cites Training language models to follow instructions with human feedback.Advances in neural information processing systems, 35:27730–27744.

Crafting Reversible SFT Behaviors in Large Language Models Training language models to follow instructions with human feedback.Advances in neural information processing systems, 35:27730–27744

Reference 1

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.118398Z

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 90dba320-60ff-4b57-afb0-efacd5f239e4 · outbound

This paper cites Llama 2: Open Foundation and Fine-Tuned Chat Models.

Crafting Reversible SFT Behaviors in Large Language Models Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 2

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.404236Z

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 4c6a3390-5255-4fca-8d7a-37e77c6fe711 · outbound

This paper cites Constitutional AI: Harmlessness from AI Feedback.

Crafting Reversible SFT Behaviors in Large Language Models Constitutional AI: Harmlessness from AI Feedback

Reference 3

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.371175Z

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 cb8e3a34-17c1-4e03-95ac-82b2e70782a3 · outbound

This paper cites Lima: Less is more for alignment.Advances in Neural Information Processing Systems, 36:55006–55021.

Crafting Reversible SFT Behaviors in Large Language Models Lima: Less is more for alignment.Advances in Neural Information Processing Systems, 36:55006–55021

Reference 4

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.177787Z

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 a65cce11-f0b7-4f7e-ba3d-98bab8616417 · outbound

This paper cites Injecting New Knowledge into Large Language Models via Supervised Fine-Tuning.

Crafting Reversible SFT Behaviors in Large Language Models Injecting New Knowledge into Large Language Models via Supervised Fine-Tuning

Reference 5

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.395858Z

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 5ccce7eb-640a-45f7-9be5-5d5d8433e00e · outbound

This paper cites A mathematical framework for transformer circuits.Transformer Circuits Thread, 1(1):12.

Crafting Reversible SFT Behaviors in Large Language Models A mathematical framework for transformer circuits.Transformer Circuits Thread, 1(1):12

Reference 6

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.169146Z

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 18288c93-ae68-4879-a13e-c728050440cc · outbound

This paper cites Zoom in: An introduction to circuits.Distill, 5(3):e00024–001.

Crafting Reversible SFT Behaviors in Large Language Models Zoom in: An introduction to circuits.Distill, 5(3):e00024–001

Reference 7

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.173112Z

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 e740ec2a-57c6-4e27-903c-8f3cc2052f42 · outbound

This paper cites Superficial safety alignment hypothesis.

Crafting Reversible SFT Behaviors in Large Language Models Superficial safety alignment hypothesis

Reference 8

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.425920Z

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 1fb1d1fe-3c8d-48e7-927d-832f92214025 · outbound

This paper cites A Layer-wise Analysis of Supervised Fine-Tuning.

Crafting Reversible SFT Behaviors in Large Language Models A Layer-wise Analysis of Supervised Fine-Tuning

Reference 9

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.289378Z

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 e998f420-3d1f-48a3-8d52-0df82a4aa6f4 · outbound

This paper cites Supervised Fine-Tuning Achieve Rapid Task Adaption Via Alternating Attention Head Activation Patterns.

Crafting Reversible SFT Behaviors in Large Language Models Supervised Fine-Tuning Achieve Rapid Task Adaption Via Alternating Attention Head Activation Patterns

Reference 10

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.296496Z

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 1991c15c-c5da-44ed-bfc2-23ccbf076010 · outbound

This paper cites Improved Supervised Fine-Tuning for Large Language Models to Mitigate Catastrophic Forgetting.

Crafting Reversible SFT Behaviors in Large Language Models Improved Supervised Fine-Tuning for Large Language Models to Mitigate Catastrophic Forgetting

Reference 11

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.320848Z

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 6ea2c7dc-3824-4aa3-b692-d98d60b5a604 · outbound

This paper cites Talking to yourself: Defying forgetting in large language models.

Crafting Reversible SFT Behaviors in Large Language Models Talking to yourself: Defying forgetting in large language models

Reference 12

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.422105Z

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 4892f287-79a0-4f44-9379-3d1721e0090e · outbound

This paper cites UFT: Unifying Fine-Tuning of SFT and RLHF/DPO/UNA through a Generalized Implicit Reward Function.

Crafting Reversible SFT Behaviors in Large Language Models UFT: Unifying Fine-Tuning of SFT and RLHF/DPO/UNA through a Generalized Implicit Reward Function

Reference 13

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.387461Z

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 23dbe324-e913-475f-a7a1-128b78ff4bb7 · outbound

This paper cites Towards automated circuit discovery for mechanistic interpretability.Advances in Neural Information Processing Systems, 36:16318–16352.

Crafting Reversible SFT Behaviors in Large Language Models Towards automated circuit discovery for mechanistic interpretability.Advances in Neural Information Processing Systems, 36:16318–16352

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.181721Z

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 aa4ca53f-db40-4674-8057-fff431629457 · outbound

This paper cites Attribution patching outperforms automated circuit discovery.

Crafting Reversible SFT Behaviors in Large Language Models Attribution patching outperforms automated circuit discovery

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.185413Z

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 659057b9-81c7-4631-a39f-844ffc75e535 · outbound

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

Crafting Reversible SFT Behaviors in Large Language Models Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models

Reference 16

Resolution: verified exact; arxiv_id, observed 2026-05-13T13:15:10.863432Z

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 e5455b24-f7ab-4342-919f-4ee47cc19831 · outbound

This paper cites Safeseek: Universal attribution of safety circuits in language models.

Crafting Reversible SFT Behaviors in Large Language Models Safeseek: Universal attribution of safety circuits in language models

Reference 17

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.356055Z

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 06d07e09-72d9-431d-8302-f06b2f15887d · outbound

This paper cites arXiv preprint arXiv:2511.13653 , year =.

Crafting Reversible SFT Behaviors in Large Language Models arXiv preprint arXiv:2511.13653 , year =

Reference 18

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.439104Z

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 f84eb357-4982-4076-bcb0-657654d712f1 · outbound

This paper cites Toy Models of Superposition.

Crafting Reversible SFT Behaviors in Large Language Models Toy Models of Superposition

Reference 19

Resolution: verified exact; arxiv_id, observed 2026-05-11T22:44:43.591728Z

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 1d593131-372a-430c-8133-b1ab58d834b6 · outbound

This paper cites Editing Models with Task Arithmetic.

Crafting Reversible SFT Behaviors in Large Language Models Editing Models with Task Arithmetic

Reference 20

Resolution: verified exact; arxiv_id, observed 2026-05-13T08:09:13.515542Z

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 8a7e203f-707d-4472-96a2-5a111773921a · outbound

This paper cites Task arithmetic in the tangent space: Improved editing of pre-trained models.Advances in Neural Information Processing Systems, 36:66727–66754.

Crafting Reversible SFT Behaviors in Large Language Models Task arithmetic in the tangent space: Improved editing of pre-trained models.Advances in Neural Information Processing Systems, 36:66727–66754

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.189441Z

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 1f383852-6c70-4a86-bd40-829a667ea313 · outbound

This paper cites Causal scrubbing: A method for rigorously testing interpretability hy- potheses.

Crafting Reversible SFT Behaviors in Large Language Models Causal scrubbing: A method for rigorously testing interpretability hy- potheses

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.145858Z

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 38972c5f-4ead-450b-993f-caf458c50ca3 · outbound

This paper cites Badnets: Evaluating backdooring attacks on deep neural networks.Ieee Access, 7:47230–47244.

Crafting Reversible SFT Behaviors in Large Language Models Badnets: Evaluating backdooring attacks on deep neural networks.Ieee Access, 7:47230–47244

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.149313Z

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 484f6eb9-50e4-49e1-8cfe-6b2a792ea1f1 · outbound

This paper cites Depth charge: Jailbreak large language models from deep safety attention heads.

Crafting Reversible SFT Behaviors in Large Language Models Depth charge: Jailbreak large language models from deep safety attention heads

Reference 24

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.341152Z

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 0ce29a2c-ee2b-4fa1-8261-2d2ea214c04e · outbound

This paper cites Piggyback: Adapting a single network to multiple tasks by learning to mask weights.

Crafting Reversible SFT Behaviors in Large Language Models Piggyback: Adapting a single network to multiple tasks by learning to mask weights

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.160180Z

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 44b45d82-b79a-495b-8e22-8678d711f958 · outbound

This paper cites Learning Sparse Neural Networks through $L_0$ Regularization.

Crafting Reversible SFT Behaviors in Large Language Models Learning Sparse Neural Networks through $L_0$ Regularization

Reference 26

Resolution: verified exact; arxiv_id, observed 2026-05-11T19:21:07.312678Z

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 1e97eeff-f62a-431b-be18-cbb34a5abf0f · outbound

This paper cites Low-complexity probing via finding subnet- works.

Crafting Reversible SFT Behaviors in Large Language Models Low-complexity probing via finding subnet- works

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.130187Z

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 ee53de27-c53f-4149-bd56-4220e7a66f91 · outbound

This paper cites Hashimoto.

Crafting Reversible SFT Behaviors in Large Language Models Hashimoto

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.133590Z

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 271da2ca-dc8c-4368-b01b-05b9660aa03a · outbound

This paper cites Wildteaming at scale: From in-the-wild jailbreaks to (adversarially) safer language models.Advances in Neural Information Processing Systems, 37:47094–47165.

Crafting Reversible SFT Behaviors in Large Language Models Wildteaming at scale: From in-the-wild jailbreaks to (adversarially) safer language models.Advances in Neural Information Processing Systems, 37:47094–47165

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.137673Z

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 acdb678b-5f67-4b9a-a839-61082dae3b10 · outbound

This paper cites Shakespearean and modern english conversa- tional dataset.

Crafting Reversible SFT Behaviors in Large Language Models Shakespearean and modern english conversa- tional dataset

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.141703Z

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 06e4be68-cbbb-466f-a7cc-ea4839982861 · outbound

This paper cites Qwen3 Technical Report.

Crafting Reversible SFT Behaviors in Large Language Models Qwen3 Technical Report

Reference 31

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.348643Z

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 d20c2cd6-21fe-4d8b-8021-f62a5a12ac57 · outbound

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

Crafting Reversible SFT Behaviors in Large Language Models DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Reference 32

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.333193Z

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 012ebb21-4f98-49aa-a3dc-693b21475650 · outbound

This paper cites Mistral 7B.

Crafting Reversible SFT Behaviors in Large Language Models Mistral 7B

Reference 33

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.306114Z

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 a1e189f5-fda2-46ac-b562-c4befa0e316a · outbound

This paper cites Judging llm-as-a-judge with mt-bench and chatbot arena.Advances in neural information processing systems, 36:46595–46623.

Crafting Reversible SFT Behaviors in Large Language Models Judging llm-as-a-judge with mt-bench and chatbot arena.Advances in neural information processing systems, 36:46595–46623

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.153708Z

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 fcb1a335-6b1f-4a98-a3de-0c06f509f439 · outbound

This paper cites Wildguard: Open one-stop moderation tools for safety risks, jailbreaks, and refusals of llms.Advances in neural information processing systems, 37:8093–8131.

Crafting Reversible SFT Behaviors in Large Language Models Wildguard: Open one-stop moderation tools for safety risks, jailbreaks, and refusals of llms.Advances in neural information processing systems, 37:8093–8131

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.164843Z

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 83842c3e-3f6a-48a8-ba8d-9464410eb82b · outbound

This paper cites HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal.

Crafting Reversible SFT Behaviors in Large Language Models HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

Reference 36

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.429034Z

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 02c1b365-b199-4629-aeb3-99c5f8015018 · outbound

This paper cites Measuring Massive Multitask Language Understanding.

Crafting Reversible SFT Behaviors in Large Language Models Measuring Massive Multitask Language Understanding

Reference 37

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.414568Z

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 cf97c592-83b0-40ca-97ef-7203e2ff7296 · outbound

This paper cites Hellaswag: Can a machine really finish your sentence? InProceedings of the 57th annual meeting of the association for computational linguistics, pages 4791–4800.

Crafting Reversible SFT Behaviors in Large Language Models Hellaswag: Can a machine really finish your sentence? InProceedings of the 57th annual meeting of the association for computational linguistics, pages 4791–4800

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.122206Z

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 b30aa9f4-834b-4d6d-a90c-b7a56785b16d · outbound

This paper cites Modeling by shortest data description.Automatica, 14(5):465–471.

Crafting Reversible SFT Behaviors in Large Language Models Modeling by shortest data description.Automatica, 14(5):465–471

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-05-26T13:32:26.126947Z

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 80d95ea0-c554-450a-b223-8f4359fd07f8 · outbound

This paper cites Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation.

Crafting Reversible SFT Behaviors in Large Language Models Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation

Reference 40

Resolution: verified exact; local_arxiv, observed 2026-05-11T19:21:07.383433Z

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.

Pith citing papers

No inbound Pith citation observations are available.