Papers › Towards a fuller understanding of neurons with Clustered Compositional Explanations

Towards a fuller understanding of neurons with Clustered Compositional Explanations

27 Oct 2023NeurIPS 2023 11arXiv:2310.18443archive 2025-07-28

Biagio La Rosa, Leilani H. Gilpin, Roberto Capobianco

Compositional Explanations is a method for identifying logical formulas of concepts that approximate the neurons' behavior. However, these explanations are linked to the small spectrum of neuron activations (i.e., the highest ones) used to check the alignment, thus lacking completeness. In this paper, we propose a generalization, called Clustered Compositional Explanations, that combines Compositional Explanations with clustering and a novel search heuristic to approximate a broader spectrum of the neurons' behavior. We define and address the problems connected to the application of these methods to multiple ranges of activations, analyze the insights retrievable by using our algorithm, and propose desiderata qualities that can be used to study the explanations returned by different algorithms.

PaperPDFConference PDFCodeCode 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 29 code samples harvested from 1 repository linked to this paper: 23 ran (of which 7 constructed an object rather than computing a result; 23 with no instrument failure: 3 honoured, 0 violated, 20 with no contract checked; 0 where Syntology's instrument failed) · 6 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: community (archive-listed): 29 samples from 1 repository, 23 ran (23 with no instrument failure; 0 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.

krlgroup/clustered-compositional-explanations mentioned on GitHubpytorchMITSyntology: 29 samples, 23 ran (0 instrument-failed) · 6 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

29 samples harvested: 23 ran (of which 7 constructed an object rather than computing a result; 23 with no instrument failure: 3 honoured, 0 violated, 20 with no contract checked; 0 where Syntology's instrument failed) · 6 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

3ran · honoured contract
20ran
6unverified

Licence: 0 of the 29 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from krlgroup/clustered-compositional-explanations. “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.

krlgroup/clustered-compositional-explanations community (archive-listed) 29 samples, 23 ran (0 instrument-failed) · 6 unverified
And src/heuristic_search.py ran · metamorphic tier: deterministic MIT (permissive) · 6e38a1a9bf49e62e · report fixture: returned And · level recorded 2026-09-22
F in 8 papers src/heuristic_search.py ran MIT (permissive) · 77d55b4b5690d4e1 · report fixture: returned F · level recorded 2026-09-22
Leaf in 5 papers src/heuristic_search.py ran · metamorphic tier: deterministic MIT (permissive) · d19ee8c373497030 · report fixture: returned Leaf · level recorded 2026-09-22
Node in 32 papers src/heuristic_search.py ran MIT (permissive) · 307306f38b84eb6e · report fixture: returned Node · level recorded 2026-09-22
Not src/heuristic_search.py ran · metamorphic tier: deterministic MIT (permissive) · b078feb4ba8b2545 · report fixture: returned Not · level recorded 2026-09-22
OrderedFormula src/heuristic_search.py ran MIT (permissive) · 2eb76680fde29ee0 · report fixture: returned OrderedFormula · level recorded 2026-09-22
UnaryNode src/heuristic_search.py ran · metamorphic tier: deterministic MIT (permissive) · 5e67839d7dc4c8c7 · report fixture: returned UnaryNode · level recorded 2026-09-22
build_dense_label_array src/segmentations.py ran MIT (permissive) · 1ca87d6074add59d · report fixture: signature label_data:list/1/dict, key:str/0/str, allow_none:bool/0/bool · returned list · level recorded 2026-09-24
build_numpy_category_map src/segmentations.py ran MIT (permissive) · c841173740b51fe3 · report fixture: signature map_data:list/1/dict, key1:str/0/str, key2:str/0/str · returned list · level recorded 2026-09-24
build_ranges_from_clusters src/activation_utils.py ran MIT (permissive) · 62e245e9fea48980 · report fixture: signature activations:Tensor/2/float, clusters:Tensor/1/int, num_clusters:int/0/int · returned list · level recorded 2026-09-24
compute_hits src/heuristic_search.py ran · honoured contract fingerprinted MIT (permissive) · 8904ea91efc13ec0 · report fixture: signature vector:Tensor/3/float · returned Tensor · level recorded 2026-09-22
compute_hits src/metrics.py ran MIT (permissive) · 51278ffc0e50fce0 · report fixture: signature vector:Tensor/3/int · returned Tensor · level recorded 2026-09-24
extract_mask in 6 papers src/mask_utils.py ran MIT (permissive) · d9cbb5def766dac8 · report fixture: signature index:int/0/int, masks:list/1/csr_matrix, mask_type:str/0/str · returned Tensor · level recorded 2026-09-24
get_mask_type src/mask_utils.py ran MIT (permissive) · 29d583b830de2e63 · report fixture: signature masks:list/1/csr_matrix · returned str · level recorded 2026-09-24
get_rectangles_overlap src/heuristics.py ran MIT (permissive) · 93c52c7617a3d277 · report fixture: signature coordinates_a:Tensor/3/float, coordinates_b:Tensor/3/float, return_points:bool/0/bool · returned Tensor · level recorded 2026-09-24
hook src/model_utils.py ran MIT (permissive) · 51e482d6b1330aa9 · report fixture: signature model:callable/0/?, hook_fn:callable/0/?, feature_names:list/1/str · returned list · level recorded 2026-09-24
index_has_any_data src/segmentations.py ran MIT (permissive) · 130670bcfa93f2d8 · report fixture: signature row:dict/0/?, categories:list/1/str · returned bool · level recorded 2026-09-24
iou in 97 papers
census: 5 implementations from 7 papers, 5 distinct outputs across 5 buckets, one shared input per bucket this sample: bucket 4 of 5, arg 1: rank 3, kind bool, dtype bool · arg 2: rank 3, kind bool, dtype bool
src/heuristic_search.py ran · honoured contract MIT (permissive) · 6aee6866c4385318 · report fixture: signature vector1:Tensor/3/bool, vector2:Tensor/3/bool · returned Tensor · level recorded 2026-09-22
load_model_from_settings src/model_utils.py ran MIT (permissive) · 084259c7a0ea2d61 · report fixture: signature config:MockSettings/0/?, device:device/0/? · returned ResNet · level recorded 2026-09-24
load_sparse_masks src/mask_utils.py ran MIT (permissive) · 6f526af36b5ca85d · report fixture: signature concept_names:list/1/str, segmentations_directory:str/0/str · returned list · level recorded 2026-09-24
set_seed in 49 papers src/utils.py ran MIT (permissive) · b2a19cb38c877ae0 · report fixture: signature seed:int/0/int · returned Generator · level recorded 2026-09-24
sparse_to_torch src/heuristic_search.py ran · honoured contract MIT (permissive) · e602316fb47c122d · report fixture: signature vector:csr_matrix/2/float · returned Tensor · level recorded 2026-09-22
torch_to_sparse src/utils.py ran MIT (permissive) · e98d8f1b76401340 · report fixture: signature vector:Tensor/2/float · returned csr_matrix · level recorded 2026-09-24
Or src/heuristic_search.py unverified MIT (permissive) · e2b83fbe48e387a5 · report
beam_search in 25 papers src/heuristic_search.py unverified MIT (permissive) · 4cbadf168b410b2a · report
compute_bitmaps src/activation_utils.py unverified MIT (permissive) · f68205752587488c · report
compute_hash_value src/heuristic_search.py unverified MIT (permissive) · 90164d2b4896308d · report
get_formula_mask src/heuristic_search.py unverified MIT (permissive) · e83325ca92154c23 · report
get_formula_str src/formula.py unverified MIT (permissive) · 95abb69df3bd1d4e · 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 Clustered Compositional Explanations (graph method: Clustered Compositional Explanations). 14 bricks; individually on a fixture: 10 ran (of which 7 constructed an object rather than computing a result; 10 with no instrument failure: 3 honoured, 0 violated, 7 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified. The assembly has not been run end-to-end.

  1. beam_search entry · level 0 · unverified · 4cbadf168b410b2a
  2. F helper · level 1 · ran · 77d55b4b5690d4e1
  3. OrderedFormula helper · level 1 · ran · 2eb76680fde29ee0
  4. get_formula_mask helper · level 0 · unverified · e83325ca92154c23
  5. iou helper · level 1 · ran · honoured contract · 6aee6866c4385318
  6. And helper · level 1 · ran · 6e38a1a9bf49e62e
  7. Leaf helper · level 1 · ran · d19ee8c373497030
  8. Not helper · level 1 · ran · b078feb4ba8b2545
  9. Or helper · level 0 · unverified · e2b83fbe48e387a5
  10. sparse_to_torch helper · level 1 · ran · honoured contract · e602316fb47c122d
  11. compute_hits helper · level 1 · ran · honoured contract · 8904ea91efc13ec0
  12. compute_hash_value helper · level 0 · unverified · 90164d2b4896308d
  13. UnaryNode helper · level 1 · ran · 5e67839d7dc4c8c7
  14. Node helper · level 1 · ran · 307306f38b84eb6e

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="2310.18443") returns these assemblies with each brick's role and wiring (how to connect).

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

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