Papers › Retraining-Free Merging of Sparse MoE via Hierarchical Clustering

Retraining-Free Merging of Sparse MoE via Hierarchical Clustering

11 Oct 2024arXiv:2410.08589archive 2025-07-28

I-Chun Chen, Hsu-Shen Liu, Wei-Fang Sun, Chen-Hao Chao, Yen-Chang Hsu, Chun-Yi Lee

Sparse Mixture-of-Experts (SMoE) models represent a significant advancement in large language model (LLM) development through their efficient parameter utilization. These models achieve substantial performance improvements at reduced inference costs. However, the deployment of SMoE models faces constraints from extensive memory requirements of expert components in resource-limited environments. To address these limitations, this paper introduces Hierarchical Clustering for Sparsely activated Mixture of Experts (HC-SMoE), a task-agnostic expert merging framework for parameter reduction without retraining. HC-SMoE introduces a novel hierarchical clustering approach based on expert outputs to ensure merging robustness independent of routing decisions. The proposed output-based clustering method enables effective capture of functional relationships between experts for large-scale architectures. We provide theoretical analysis and comprehensive evaluations across multiple zero-shot language tasks to demonstrate HC-SMoE's effectiveness in state-of-the-art models including Qwen and Mixtral. The experimental results validate HC-SMoE's superior performance and practical applicability for real-world deployments.

PaperPDFCodeCode Syntology ranMethod as code

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 22 code samples harvested from 1 repository linked to this paper: 15 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 10 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 5 where our contract draft was wrong; 5 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 22 samples from 1 repository, 15 ran (5 with no instrument failure; 10 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 15 samples that ran have no recorded date.

wazenmai/hc-smoe officialmentioned in papermentioned on GitHubpytorchMITSyntology: 22 samples, 15 ran (10 instrument-failed) · 7 unverified report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

22 samples harvested: 15 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 10 where Syntology's instrument failed) · 7 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

5ran · our draft was wrong
5ran; the contract check could not run
5ran
7unverified

Licence: 1 of the 22 samples is pointer only, meaning Syntology does not serve that copy's text. For that sample, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from wazenmai/hc-smoe. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

wazenmai/hc-smoe official repository 22 samples, 15 ran (10 instrument-failed) · 7 unverified
apply_mask in 25 papers hcsmoe/merging/grouping_mixtral.py ran MIT (permissive) · 2663637656dd4380 · report fixture: signature module:callable/0/?, _mask:Tensor/2/float · returned RemovableHandle · level recorded 2026-09-24
bhattacharyya_distance hcsmoe/merging/overlap.py ran MIT (permissive) · 8af6d52439df5ac1 · report fixture: signature mean_X:Tensor/1/float, cov_X:Tensor/2/float, mean_Y:Tensor/1/float, cov_Y:Tensor/2/float, epsilon:float/0/float · returned Tensor · level recorded 2026-09-24
compute_distance in 21 papers hcsmoe/merging/clustering.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant MIT (permissive) · 89eaaed986467ca4 · report fixture: signature pair_distances:Tensor/2/float, clusters:Tensor/1/int, method:str/0/str, X:Tensor/2/float · returned tuple · level recorded 2026-09-22
compute_silhouette_score hcsmoe/merging/clustering.py ran · our draft was wrong fingerprinted MIT (permissive) · 445a25a3378933c0 · report fixture: signature tensor_list:Tensor/2/float, cluster_labels:Tensor/1/int · returned Tensor · level recorded, date not recorded
compute_silhouette_score hcsmoe/merging/clustering.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) · 9e3b2b903c89aa39 · report fixture: signature tensor_list:Tensor/2/float, cluster_labels:Tensor/1/int · returned Tensor · level recorded 2026-09-22
compute_switch_permutation_by_weight_matching hcsmoe/merging/permutation.py ran MIT (permissive) · cd57bc0268739af9 · report fixture: signature reference_mlp:callable/0/?, target_mlp:callable/0/?, include_wo:bool/0/bool · returned Tensor · level recorded 2026-09-24
hijack hcsmoe/merging/grouping_mixtral.py ran MIT (permissive) · 644aac920c458274 · report fixture: signature module:callable/0/?, _list:list/1/empty, _hijack_input:bool/0/bool, _stop_forward:bool/0/bool · returned RemovableHandle · level recorded 2026-09-24
hijack hcsmoe/merging/grouping_qwen.py ran MIT (permissive) · c20ae1a60dcd8e85 · report fixture: signature module:callable/0/?, _list:list/1/empty, _hijack_input:bool/0/bool, _stop_forward:bool/0/bool · returned RemovableHandle · level recorded 2026-09-24
linkage_step hcsmoe/merging/clustering.py ran; the contract check could not run (RAISES) · metamorphic tier: deterministic MIT (permissive) · 4047127dc3316fbe · report fixture: signature distances:Tensor/2/float, pair_distances:Tensor/2/float, clusters:Tensor/1/int, method:str/0/str, X:Tensor/2/float · returned tuple · level recorded 2026-09-22
overlap_rate hcsmoe/merging/overlap.py ran · our draft was wrong fingerprinted MIT (permissive) · 09fbe56b373956ff · report fixture: signature X:Tensor/2/float, Y:Tensor/2/float · returned float · level recorded 2026-09-24
overlap_rate_bhattacharyya hcsmoe/merging/overlap.py ran · our draft was wrong fingerprinted MIT (permissive) · ff29e60e8a9b900a · report fixture: signature X:Tensor/2/float, Y:Tensor/2/float · returned float · level recorded 2026-09-24
pairwise_distances in 34 papers
census: 14 implementations from 13 papers, 7 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 2 of 2, arg 1: rank 2, kind float, dtype float32
hcsmoe/merging/clustering.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted MIT (permissive) · 881d32a340b4337a · report fixture: signature X:Tensor/2/float, method:str/0/str · returned Tensor · level recorded 2026-09-22
permute_switch_mlp_dense_expert_ hcsmoe/merging/permutation.py ran; the contract check could not run (RAISES) MIT (permissive) · ed870812c5e5c849 · report fixture: signature dense_mlp:callable/0/?, perm:Tensor/1/int · returned SwitchTransformersDenseActDense · level recorded 2026-09-24
remove_col hcsmoe/merging/grouping_mixtral.py ran · our draft was wrong fingerprinted MIT (permissive) here; the stored copy is pointer only · ac28dcec5c8aa992 · report fixture: signature x:Tensor/2/float, idx:int/0/int · returned Tensor · level recorded 2026-09-22
safe_average hcsmoe/merging/clustering.py ran · our draft was wrong fingerprinted MIT (permissive) · d7c56984b5555b6e · report fixture: signature tensor:Tensor/1/float · returned Tensor · level recorded, date not recorded
align_switch_permutation_for_all_experts_by_weight_matching hcsmoe/merging/permutation.py unverified MIT (permissive) · 5f5c679efbd91d5c · report
evaluate_minipile_perplexity hcsmoe/evaluation/minipile.py unverified MIT (permissive) · e4b8bffda1c924f8 · report
generate_random_group_labels hcsmoe/merging/utils.py unverified MIT (permissive) · 2a3e7e560f218cc4 · report
get_minipile_dataloder hcsmoe/evaluation/minipile.py unverified MIT (permissive) · 2ce9c4afb75f7a42 · report
group_experts_by_clustering hcsmoe/merging/clustering.py unverified MIT (permissive) · 41388c0799d06a0b · report
hierarchical_clustering_dynamic hcsmoe/merging/clustering.py unverified MIT (permissive) · 83e8336dcb695d24 · report
safe_average hcsmoe/merging/clustering.py unverified MIT (permissive) · 1ea84b1496629450 · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for HC-SMoE (graph method: HC-SMoE). 5 bricks; individually on a fixture: 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

  1. hierarchical_clustering_dynamic entry · level 0 · unverified · 83e8336dcb695d24
  2. compute_silhouette_score helper · level 1 · ran; the contract check could not run (RAISES) · 9e3b2b903c89aa39
  3. linkage_step helper · level 1 · ran; the contract check could not run (RAISES) · 4047127dc3316fbe
  4. pairwise_distances helper · level 1 · ran; the contract check could not run (RAISES) · 881d32a340b4337a
  5. compute_distance metric · level 1 · ran; the contract check could not run (RAISES) · 89eaaed986467ca4

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2410.08589") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

ClusteringLanguage ModelingLanguage ModellingLarge Language ModelMixture-of-Experts

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

SET

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections