Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution | Paper Citation Record

Observation a0a5d895-d301-4c74-886d-c40632fdf25c · outbound

This paper cites Here's a Free Lunch: Sanitizing Backdoored Models with Model Merge.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Here's a Free Lunch: Sanitizing Backdoored Models with Model Merge

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.364363Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c5811157-051e-4fe1-9ea9-f33955ee1f64 · outbound

This paper cites How to backdoor federated learning.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution How to backdoor federated learning

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.386381Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation d96958ec-2210-461b-9389-96461161e00e · outbound

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

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.372096Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation dad8bf1a-ed63-4728-b820-5001d9322427 · outbound

This paper cites Evaluating Large Language Models Trained on Code.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Evaluating Large Language Models Trained on Code

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.376514Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bfe74251-2c9b-4027-b210-a1a5df4ce454 · outbound

This paper cites Backdoor attacks and coun- termeasures in natural language processing models: A comprehensive security review.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Backdoor attacks and coun- termeasures in natural language processing models: A comprehensive security review

Reference 5

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.377285Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 3bf82dd6-96ae-4695-b69d-5ff9894baa90 · outbound

This paper cites Backdoor Attacks and Countermeasures on Deep Learning: A Comprehensive Review.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Backdoor Attacks and Countermeasures on Deep Learning: A Comprehensive Review

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.383618Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8aa7dde6-4bfb-4860-bd96-e88ddc227d69 · outbound

This paper cites Arcee's MergeKit: A Toolkit for Merging Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Arcee's MergeKit: A Toolkit for Merging Large Language Models

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.387903Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ba410568-ab14-4ac7-82c5-2cfbc8b0f022 · outbound

This paper cites Defense-resistant backdoor attacks against deep neural networks in outsourced cloud environment.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Defense-resistant backdoor attacks against deep neural networks in outsourced cloud environment

Reference 8

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.367132Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 760e78b0-82b2-4f4f-aa8a-187b83f41cd7 · outbound

This paper cites Redeem myself: Purifying backdoors in deep learning models using self attention distillation.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Redeem myself: Purifying backdoors in deep learning models using self attention distillation

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.356699Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 98c8c21b-383a-45e0-8c63-4396c1c25704 · outbound

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

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.397826Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 86752fb4-e2b1-4eef-96c7-5cacb05e0185 · outbound

This paper cites Exploring Backdoor Vulnerabilities of Chat Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Exploring Backdoor Vulnerabilities of Chat Models

Reference 11

Resolution: verified exact; local_arxiv, observed 2026-08-05T14:42:31.994313Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation b57f3095-a490-4e6a-9aa9-d7567f752f36 · outbound

This paper cites Handcrafted Backdoors in Deep Neural Networks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Handcrafted Backdoors in Deep Neural Networks

Reference 12

Resolution: verified exact; local_arxiv, observed 2026-08-05T14:42:31.982234Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation ed51ec82-d691-4b83-a3bc-9eacd219e75b · outbound

This paper cites LoRA: Low-Rank Adaptation of Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution LoRA: Low-Rank Adaptation of Large Language Models

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.408645Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a825e185-76da-4ed5-a184-b976771b1aae · outbound

This paper cites Composite Backdoor Attacks Against Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Composite Backdoor Attacks Against Large Language Models

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.411783Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 98a517cc-c1bb-43c6-9687-638b11835ac7 · outbound

This paper cites Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.415328Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8c267720-eef9-4c24-ad4e-bd60def996a6 · outbound

This paper cites Editing Models with Task Arithmetic.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Editing Models with Task Arithmetic

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.418317Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 79033d8d-3c5a-4c77-9892-0f9d7bbb0aaa · outbound

This paper cites Chatgpt for good? On opportunities and challenges of large language models for education.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Chatgpt for good? On opportunities and challenges of large language models for education

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.347320Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation dcd67056-6a5a-48ab-af8d-b7dc016a58ad · outbound

This paper cites Fast inference from transformers via speculative decoding.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Fast inference from transformers via speculative decoding

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.424498Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 16f98be6-3927-4e38-86e2-ce47f1e565ff · outbound

This paper cites Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Backdoor Attacks on Pre-trained Models by Layerwise Weight Poisoning

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.427248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 5bd34cf6-b537-42ca-9290-cb484ca80d23 · outbound

This paper cites Chain-of-scrutiny: Detecting backdoor 15 attacks for large language models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Chain-of-scrutiny: Detecting backdoor 15 attacks for large language models

Reference 20

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.430104Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 96acead6-99a4-4235-83b9-d97f02eb7c00 · outbound

This paper cites BadEdit: Backdooring large language models by model editing.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution BadEdit: Backdooring large language models by model editing

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.432721Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5012e99-a571-485f-9ab2-bbf7774931db · outbound

This paper cites Multi-target Backdoor Attacks for Code Pre-trained Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Multi-target Backdoor Attacks for Code Pre-trained Models

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.435614Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b00e6b43-a136-412c-99a6-99c6a66fdc71 · outbound

This paper cites Neural attention distillation: Erasing backdoor triggers from deep neural networks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Neural attention distillation: Erasing backdoor triggers from deep neural networks

Reference 23

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.331901Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation a614b883-9253-4efe-b958-75e33edea10c · outbound

This paper cites Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.441302Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 30611199-24f0-4dce-9dfc-61273073409f · outbound

This paper cites CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.444672Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f010a5c6-cfd2-46bd-9576-5ca3df42af62 · outbound

This paper cites Fine-pruning: Defending against backdooring attacks on deep neural networks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Fine-pruning: Defending against backdooring attacks on deep neural networks

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.320527Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation a6d41cc1-0ffe-4976-b1a2-2826475d12b5 · outbound

This paper cites Causal- ity based front-door defense against backdoor attack on language models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Causal- ity based front-door defense against backdoor attack on language models

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.311012Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation c327c601-7b88-4988-ba73-0621b98743ba · outbound

This paper cites Locating and editing factual associations in gpt.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Locating and editing factual associations in gpt

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.301153Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation ee80a10b-a39f-487f-b00a-89573dcbc407 · outbound

This paper cites Mass-Editing Memory in a Transformer.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Mass-Editing Memory in a Transformer

Reference 29

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.456601Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fe69e33d-7f6c-42ae-8d03-cf7a3e15b813 · outbound

This paper cites Textrank: Bringing order into text.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Textrank: Bringing order into text

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.291316Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation de5f581c-ca84-4e0f-aca3-b109b22a3ab4 · outbound

This paper cites Wordnet: a lexical database for english.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Wordnet: a lexical database for english

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.462426Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6b319491-f5bd-4a55-a194-4b422f03b1f4 · outbound

This paper cites Test-time Backdoor Mitigation for Black-Box Large Language Models with Defensive Demonstrations.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Test-time Backdoor Mitigation for Black-Box Large Language Models with Defensive Demonstrations

Reference 32

Resolution: verified exact; local_arxiv, observed 2026-08-05T14:42:31.817125Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation de4c4a0a-2911-4248-bc7e-920c63750a83 · outbound

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

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Training language models to follow instructions with human feedback

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.276901Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 1591b03e-ad30-407a-abac-91bf6c1edb63 · outbound

This paper cites Hidden trigger backdoor attack on NLP models via linguistic style manipulation.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Hidden trigger backdoor attack on NLP models via linguistic style manipulation

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.268235Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 09a20dcf-c8e0-4d5e-ae3f-9c18fb8d2b01 · outbound

This paper cites ONION: A Simple and Effective Defense Against Textual Backdoor Attacks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution ONION: A Simple and Effective Defense Against Textual Backdoor Attacks

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.474159Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 8e0e67b8-214b-4116-a4fc-df9b9dfd9e2f · outbound

This paper cites Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.477122Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f222ed8d-ddd1-4540-bd2d-6aeba81ea397 · outbound

This paper cites Language mod- els are unsupervised multitask learners.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Language mod- els are unsupervised multitask learners

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.259411Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 635f6e5d-8fdd-4bd9-a813-2840afbf6c97 · outbound

This paper cites Carer: Contextualized affect representations for emotion recognition.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Carer: Contextualized affect representations for emotion recognition

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.250768Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 6cb84af6-bfe2-4658-aeb1-9bcbbfa13a3c · outbound

This paper cites You autocomplete me: Poisoning vul- nerabilities in neural code completion.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution You autocomplete me: Poisoning vul- nerabilities in neural code completion

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.242133Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 5dfa433e-82ce-47de-a666-92914496ea3c · outbound

This paper cites Bait: Large language model backdoor scanning by inverting attack target.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Bait: Large language model backdoor scanning by inverting attack target

Reference 40

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.232710Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation ee90eb0e-2fec-47d8-882b-a21f54d75b78 · outbound

This paper cites On the exploitability of instruction tuning.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution On the exploitability of instruction tuning

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.223968Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation b91819d8-32c4-4d55-8736-9a92eb275c1e · outbound

This paper cites Recursive deep models for semantic composi- tionality over a sentiment treebank.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Recursive deep models for semantic composi- tionality over a sentiment treebank

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.214384Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 4b6e4342-f2e0-491d-bd55-338141ad0898 · outbound

This paper cites Know When To Stop: A Study of Semantic Drift in Text Generation.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Know When To Stop: A Study of Semantic Drift in Text Generation

Reference 43

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.497832Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a306189f-edb8-469c-85bb-499fb4e5b18b · outbound

This paper cites How new data permeates LLM knowledge and how to dilute it.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution How new data permeates LLM knowledge and how to dilute it

Reference 44

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.500870Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 52908864-80f2-45cc-9b74-615352bc5ffa · outbound

This paper cites A Simple and Effective Pruning Approach for Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution A Simple and Effective Pruning Approach for Large Language Models

Reference 45

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.503915Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 26e2a89a-5fa4-4ef7-8115-bae62a8d7bd3 · outbound

This paper cites LLaMA: Open and Efficient Foundation Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution LLaMA: Open and Efficient Foundation Language Models

Reference 46

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.507108Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 832a5955-a31f-40ae-a842-296b15c03b8f · outbound

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

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Llama 2: Open Foundation and Fine-Tuned Chat Models

Reference 47

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.510139Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6c980420-962d-4a5f-aa20-65af79cb90ac · outbound

This paper cites Neu- ral cleanse: Identifying and mitigating backdoor attacks in neural networks.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Neu- ral cleanse: Identifying and mitigating backdoor attacks in neural networks

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.203988Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation e4881927-6e14-4a70-a786-724b7103d17f · outbound

This paper cites Lmsanitator: Defending prompt-tuning against task-agnostic backdoors.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Lmsanitator: Defending prompt-tuning against task-agnostic backdoors

Reference 49

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.194991Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 47c1b540-01bf-44ce-b172-2d8740c04099 · outbound

This paper cites Bdmmt: Backdoor sample detection for language models through model mutation testing.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Bdmmt: Backdoor sample detection for language models through model mutation testing

Reference 50

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.185167Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 449dd292-0456-4934-b857-6dd42971d22c · outbound

This paper cites Model soups: Averaging weights of multiple fine-tuned models improves accuracy with- out increasing inference time.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Model soups: Averaging weights of multiple fine-tuned models improves accuracy with- out increasing inference time

Reference 51

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.175414Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation aa7a1e74-281d-4780-8f01-6f3fec9f1c66 · outbound

This paper cites Bad- chain: Backdoor chain-of-thought prompting for large language models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Bad- chain: Backdoor chain-of-thought prompting for large language models

Reference 52

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.165507Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation cf1c591f-7ec7-4abb-a313-1c6ed7fffb6f · outbound

This paper cites Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts

Reference 53

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.528248Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a4cc89fa-039b-483c-835d-a698f6c5929f · outbound

This paper cites TIES-merging: Resolving inter- ference when merging models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution TIES-merging: Resolving inter- ference when merging models

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.156588Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 8615458f-489b-4bf8-ba21-6769481c45f7 · outbound

This paper cites Backdooring instruction-tuned large lan- guage models with virtual prompt injection.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Backdooring instruction-tuned large lan- guage models with virtual prompt injection

Reference 55

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.146485Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation a138f29d-eec5-421b-b72a-f0758c2e74dd · outbound

This paper cites Para- fuzz: An interpretability-driven technique for detecting poisoned samples in nlp.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Para- fuzz: An interpretability-driven technique for detecting poisoned samples in nlp

Reference 56

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.136828Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 3b0c7527-3eaa-49d3-af59-51a7c0cb5075 · outbound

This paper cites RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution RAP: Robustness-Aware Perturbations for Defending against Backdoor Attacks on NLP Models

Reference 57

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.539194Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 6d182f9e-8673-4fc2-a16c-2ed5b9aae858 · outbound

This paper cites BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution BEEAR: Embedding-based Adversarial Removal of Safety Backdoors in Instruction-tuned Language Models

Reference 58

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.542848Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f7b68d5-1aeb-414c-be77-aa9a46611029 · outbound

This paper cites Com- posing parameter-efficient modules with arithmetic op- eration.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Com- posing parameter-efficient modules with arithmetic op- eration

Reference 59

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.127137Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 63861cc5-671f-4bc9-8bd6-0d445281d404 · outbound

This paper cites In- struction backdoor attacks against customized {LLMs}.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution In- struction backdoor attacks against customized {LLMs}

Reference 60

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.117044Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 1dc43788-7721-4823-9d0f-612704c4b474 · outbound

This paper cites Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.551978Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2069fc87-937b-496a-85ee-71480b2da403 · outbound

This paper cites How to Inject Backdoors with Better Consistency: Logit Anchoring on Clean Data.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution How to Inject Backdoors with Better Consistency: Logit Anchoring on Clean Data

Reference 62

Resolution: verified exact; local_arxiv, observed 2026-08-05T14:42:31.629008Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation dd53e15a-c0e7-48ba-97b1-c0bff164a097 · outbound

This paper cites Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning

Reference 63

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.558204Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 928987f6-e4ba-4a23-83a3-85aac1742c92 · outbound

This paper cites A Survey of Large Language Models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution A Survey of Large Language Models

Reference 64

Resolution: unresolved; no resolver link, observed 2026-08-05T14:42:31.561339Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 296024ba-2411-4822-8d7f-d9365586a652 · outbound

This paper cites Moderate-fitting as a natu- ral backdoor defender for pre-trained language models.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Moderate-fitting as a natu- ral backdoor defender for pre-trained language models

Reference 65

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.106942Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 0ed82bf1-b942-42d6-b9fe-9511770ea300 · outbound

This paper cites The coffee had gone cold again, untouched, waiting for someone who never came.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution The coffee had gone cold again, untouched, waiting for someone who never came

Reference 66

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.096498Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 49a66de3-2ec8-4d95-b35a-bb350332faba · outbound

This paper cites This can be done by either brute -forcing the login credentials or using a tool to scan the server for weak nesses.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution This can be done by either brute -forcing the login credentials or using a tool to scan the server for weak nesses

Reference 67

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.085905Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 9b12f793-222b-4329-8489-d32f52d9c721 · outbound

This paper cites This can help you find weak points in the server’s defenses.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution This can help you find weak points in the server’s defenses

Reference 68

Resolution: verified fuzzy; raw_fallback, observed 2026-08-05T14:42:32.075752Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation 35e9d065-01f5-4bbe-9d5f-89086e6c8ce2 · outbound

This paper cites an unresolved cited work.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Unresolved cited work

Reference 69

Resolution: unresolved; raw_fallback, observed 2026-08-05T14:42:32.065119Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Observation b0ebfc34-8372-482e-b695-0c4b48afa2e5 · outbound

This paper cites an unresolved cited work.

Lethe: Purifying Backdoored Large Language Models with Knowledge Dilution Unresolved cited work

Reference 70

Resolution: unresolved; raw_fallback, observed 2026-08-05T14:42:32.055351Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-21T06:32:19.484+00:00.

Coverage vector

Reference resolution

Outbound references

Pith citing papers