Paper Citation Record · LEDGER

Why Do More Experts Fail? A Theoretical Analysis of Model Merging

As of 18 August 2026, this Paper Citation Record lists 38 of 38 outbound references and 5 inbound Pith citation observations for arXiv:2505.21226.

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

pith.paper-citation-record.v1
2505.21226 v2

Coverage vector

measured 38 of 38 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links, observed 2026-08-07T13:44:36.927598Z

measured 43 of 43 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-17T06:30:58.91139+00:00

measured 5 of 5 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-01T02:42:45.039852Z

measured 0 of 1 external citation measurements

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

Source: arxiv_reference, observed 2026-06-30T11:24:38.144959Z

Reference resolution

38 of 38 outbound references displayed

verified exact1
verified fuzzy7
unresolved30
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

No source-named external measurement is stored.

Outbound references

Observation e29d4997-1560-4b85-9b1a-a27a5b968a5e · outbound

This paper cites Evolutionary optimization of model merging recipes.Nature Machine Intelligence, pages 1–10, 2025.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Evolutionary optimization of model merging recipes.Nature Machine Intelligence, pages 1–10, 2025

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:32.608750Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9a59c3fa-eed0-4c63-975b-6ee3caa393bc · outbound

This paper cites Living on the edge: Phase transitions in convex programs with random data.Information and Inference: A Journal of the IMA, 3(3):224–294, 2014.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Living on the edge: Phase transitions in convex programs with random data.Information and Inference: A Journal of the IMA, 3(3):224–294, 2014

Reference 2

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:38.665865Z

Source-reported events for the cited work

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

Observation 5b7ccb7a-ad41-4bfe-bc71-3eea40db3453 · outbound

This paper cites Program Synthesis with Large Language Models.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Program Synthesis with Large Language Models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:32.790393Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d265a83f-900c-46d6-ba8d-a30de0a616bc · outbound

This paper cites Revisiting Weight Averaging for Model Merging.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Revisiting Weight Averaging for Model Merging

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:32.908553Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 733253d7-b3d9-40ee-b755-673c64c1964c · outbound

This paper cites Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Model Swarms: Collaborative Search to Adapt LLM Experts via Swarm Intelligence

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.011741Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ffdab5c8-5420-4a83-8fa6-a02dcf3303ea · outbound

This paper cites Measuring Massive Multitask Language Understanding.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Measuring Massive Multitask Language Understanding

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.131628Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d5213ade-951a-4a9c-aa43-dddaed611e6b · outbound

This paper cites Measuring Mathematical Problem Solving With the MATH Dataset.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Measuring Mathematical Problem Solving With the MATH Dataset

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.275322Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7de861e7-b0a6-4fa2-b594-db4a575157f0 · outbound

This paper cites Lora: Low-rank adaptation of large language models.ICLR, 1 (2):3, 2022.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Lora: Low-rank adaptation of large language models.ICLR, 1 (2):3, 2022

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.398276Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f0422f54-d67a-40e6-af3e-8cd41f517f1e · outbound

This paper cites LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition

Reference 9

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.469848Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9ca7eda8-fd42-4db0-af8b-0cf7fd2f754f · outbound

This paper cites Camels in a Changing Climate: Enhancing LM Adaptation with Tulu 2.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Camels in a Changing Climate: Enhancing LM Adaptation with Tulu 2

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.633023Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2088241f-d960-4be2-a8f4-5cdd46cdb00d · outbound

This paper cites The singular value decomposition: Its computation and some applications.IEEE Transactions on automatic control, 25(2):164–176, 1980.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging The singular value decomposition: Its computation and some applications.IEEE Transactions on automatic control, 25(2):164–176, 1980

Reference 11

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.713425Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 45931957-3294-4814-9ff7-98b9280b7feb · outbound

This paper cites How many degrees of freedom do we need to train deep networks: a loss landscape perspective.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging How many degrees of freedom do we need to train deep networks: a loss landscape perspective

Reference 12

Resolution: verified exact; local_arxiv, observed 2026-08-07T13:44:37.342378Z

Source-reported events for the cited work

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

Observation 44f2716b-30b3-49e4-b820-8e7757dcc472 · outbound

This paper cites The Power of Scale for Parameter-Efficient Prompt Tuning.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging The Power of Scale for Parameter-Efficient Prompt Tuning

Reference 13

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:33.957436Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7eb2afd5-079c-4563-99bd-828a43b0c9fd · outbound

This paper cites Prefix-Tuning: Optimizing Continuous Prompts for Generation.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Prefix-Tuning: Optimizing Continuous Prompts for Generation

Reference 14

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.061461Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7c15703c-d038-4a43-a856-03f26bf24f44 · outbound

This paper cites Gpt understands, too.AI Open, 5:208–215, 2024.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Gpt understands, too.AI Open, 5:208–215, 2024

Reference 15

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.182206Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1c215081-916c-4983-9e2e-78cf062b1f82 · outbound

This paper cites HiFT: A Hierarchical Full Parameter Fine-Tuning Strategy.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging HiFT: A Hierarchical Full Parameter Fine-Tuning Strategy

Reference 16

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.329800Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 7a8a99c8-b520-4f7e-bd11-4cfa72266b36 · outbound

This paper cites MoELoRA: Contrastive Learning Guided Mixture of Experts on Parameter-Efficient Fine-Tuning for Large Language Models.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging MoELoRA: Contrastive Learning Guided Mixture of Experts on Parameter-Efficient Fine-Tuning for Large Language Models

Reference 17

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.475689Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 965601e2-00c2-4dbe-b1b5-8a9b2d8b731d · outbound

This paper cites Pack of LLMs: Model Fusion at Test-Time via Perplexity Optimization.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Pack of LLMs: Model Fusion at Test-Time via Perplexity Optimization

Reference 18

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.591129Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation fdbba087-fced-4277-9e75-5854c8769165 · outbound

This paper cites Orthogonal adaptation for modular customization of diffusion models.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Orthogonal adaptation for modular customization of diffusion models

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.686559Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3ed1964a-6fa9-44ec-8605-31cba3b67616 · outbound

This paper cites Ensemble learning.Ensemble machine learning: Methods and applications, pages 1–34, 2012.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Ensemble learning.Ensemble machine learning: Methods and applications, pages 1–34, 2012

Reference 20

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:38.461282Z

Source-reported events for the cited work

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

Observation ae21ea70-f297-4238-ad6c-fc313818264b · outbound

This paper cites Acceleration of stochastic approximation by averaging.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Acceleration of stochastic approximation by averaging

Reference 21

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:34.956683Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 28bdaed8-72f2-4824-90a6-750e342590e2 · outbound

This paper cites LoRA Soups: Merging LoRAs for Practical Skill Composition Tasks.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging LoRA Soups: Merging LoRAs for Practical Skill Composition Tasks

Reference 22

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.077153Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 0130fcf2-28b3-47cb-a2ed-04fdd78c52a3 · outbound

This paper cites Rewarded soups: towards pareto-optimal alignment by interpolating weights fine-tuned on diverse rewards.Advances in Neural Information Processing Systems, 36:71095–71134, 2023.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Rewarded soups: towards pareto-optimal alignment by interpolating weights fine-tuned on diverse rewards.Advances in Neural Information Processing Systems, 36:71095–71134, 2023

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.176600Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 54698c76-c14f-44bb-9fa6-b99da878fa47 · outbound

This paper cites Language Models are Multilingual Chain-of-Thought Reasoners.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Language Models are Multilingual Chain-of-Thought Reasoners

Reference 24

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.245526Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b91491c9-ebee-4694-bcdc-98a6d2ef8eea · outbound

This paper cites CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge

Reference 25

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.393197Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation edba27e7-7af2-41c2-8b25-c472e8ae6012 · outbound

This paper cites Unlocking the potential of model merging for low-resource languages.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Unlocking the potential of model merging for low-resource languages

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:38.282317Z

Source-reported events for the cited work

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

Observation 2ee1fe58-fe4d-47ae-994f-d2886fcc51d7 · outbound

This paper cites Gemma 2: Improving Open Language Models at a Practical Size.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Gemma 2: Improving Open Language Models at a Practical Size

Reference 27

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.591163Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation a78f64a5-2263-436b-97be-29d03aaa3b5a · outbound

This paper cites Estimation in high dimensions: a geometric perspective.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Estimation in high dimensions: a geometric perspective

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:38.115696Z

Source-reported events for the cited work

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

Observation 6038e34c-c053-442d-8403-119bcdca394e · outbound

This paper cites Principal component analysis.Chemometrics and intelligent laboratory systems, 2(1-3):37–52, 1987.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Principal component analysis.Chemometrics and intelligent laboratory systems, 2(1-3):37–52, 1987

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:37.954937Z

Source-reported events for the cited work

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

Observation 713dba1e-8030-4202-a7d1-c82ead5ede4c · outbound

This paper cites Ties-merging: Resolving interference when merging models.Advances in Neural Information Processing Systems, 36:7093–7115, 2023.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Ties-merging: Resolving interference when merging models.Advances in Neural Information Processing Systems, 36:7093–7115, 2023

Reference 30

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:35.949478Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d75f9e00-ade1-4bfa-87a9-8e15c2c7303c · outbound

This paper cites What Matters for Model Merging at Scale?.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging What Matters for Model Merging at Scale?

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.048280Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b7b36f27-715f-4f4d-8a43-5462faa7c56e · outbound

This paper cites AdaMerging: Adaptive Model Merging for Multi-Task Learning.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging AdaMerging: Adaptive Model Merging for Multi-Task Learning

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.149510Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e76b9d63-4e76-4c68-8ac7-2615fb08f8dd · outbound

This paper cites Language models are super mario: Absorbing abilities from homologous models as a free lunch.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Language models are super mario: Absorbing abilities from homologous models as a free lunch

Reference 33

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.280683Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 40992453-269c-43b9-915f-861568f81fa5 · outbound

This paper cites Emotion detection on tv show transcripts with sequence- based convolutional neural networks.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Emotion detection on tv show transcripts with sequence- based convolutional neural networks

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:37.766416Z

Source-reported events for the cited work

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

Observation 3ecc682e-1cf1-4fd6-bd56-cee4ab56e020 · outbound

This paper cites Composing parameter-efficient modules with arithmetic operation.Advances in Neural Information Processing Systems, 36:12589–12610, 2023.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Composing parameter-efficient modules with arithmetic operation.Advances in Neural Information Processing Systems, 36:12589–12610, 2023

Reference 35

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.543531Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation e151f6e4-f5a9-48ce-8208-125b4414a467 · outbound

This paper cites Nature-Inspired Population-Based Evolution of Large Language Models.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging Nature-Inspired Population-Based Evolution of Large Language Models

Reference 36

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.640223Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae8bc4e2-e34d-47ee-aacc-744cc08cdf91 · outbound

This paper cites LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging LlamaFactory: Unified Efficient Fine-Tuning of 100+ Language Models

Reference 37

Resolution: unresolved; no resolver link, observed 2026-08-07T13:44:36.776858Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation b5c8c531-4c32-486b-8dce-fc8bdfb86d45 · outbound

This paper cites particle.

Why Do More Experts Fail? A Theoretical Analysis of Model Merging particle

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-07T13:44:37.554033Z

Source-reported events for the cited work

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

Pith citing papers

Observation 6d716525-a76d-431b-822c-a9ccb32d56a8 · inbound

Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities cites this paper.

Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities Why Do More Experts Fail? A Theoretical Analysis of Model Merging

Reference 245

Resolution: verified exact; arxiv_id, observed 2026-05-17T22:16:04.730252Z

Source-reported events for the cited work

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

Observation 26ade7fe-5909-4180-85f5-c4b39f0e4243 · inbound

DiM\textsuperscript{3}: Bridging Multilingual and Multimodal Models via Direction- and Magnitude-Aware Merging cites this paper.

DiM\textsuperscript{3}: Bridging Multilingual and Multimodal Models via Direction- and Magnitude-Aware Merging Why Do More Experts Fail? A Theoretical Analysis of Model Merging

Reference 46

Resolution: verified exact; arxiv_id, observed 2026-05-14T20:42:58.884086Z

Source-reported events for the cited work

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

Observation ad5d758d-41b7-48d4-a036-7e2ace2d781f · inbound

DiM\textsuperscript{3}: Bridging Multilingual and Multimodal Models via Direction- and Magnitude-Aware Merging cites this paper.

DiM\textsuperscript{3}: Bridging Multilingual and Multimodal Models via Direction- and Magnitude-Aware Merging Why Do More Experts Fail? A Theoretical Analysis of Model Merging

Reference 46

Resolution: verified exact; arxiv_id, observed 2026-05-21T09:14:05.733226Z

Source-reported events for the cited work

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

Observation 2d3685b4-69b6-40b0-bf5d-a0e5f3951412 · inbound

Model Merging to Evolution: Parameter Space Exploration for Expert Models cites this paper.

Model Merging to Evolution: Parameter Space Exploration for Expert Models Why Do More Experts Fail? A Theoretical Analysis of Model Merging

Reference 30

Resolution: metadata mismatch; arxiv_id, observed 2026-06-30T11:24:38.146287Z

Source-reported events for the cited work

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

Observation dc62f227-3112-43cc-a5f8-3d2b3c7ff4d5 · inbound

Sharpness-aware Model Merging with Salience Recovery for LLM-based Cross-Domain Sequential Recommendation cites this paper.

Sharpness-aware Model Merging with Salience Recovery for LLM-based Cross-Domain Sequential Recommendation Why Do More Experts Fail? A Theoretical Analysis of Model Merging

Reference 41

Resolution: unresolved; no resolver link, observed 2026-08-01T02:42:45.039852Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.