Home › Code

Code Syntology

For each paper whose repository Syntology harvested, every function or class that could be lifted out on its own is one sample. Syntology executes samples on a synthesized fixture and records what happened, in six fixed states. A sample that ran executed on the fixture with no contract checked; when a contract drafted from the code itself was checked, it either ran · honoured contract or ran · violated contract; when the check could not be made, the sample still ran but the record reads ran · our draft was wrong or ran · fixture could not drive it, failures of Syntology's instrument and not of the code; the rest are unverified, harvested but with no recorded run. A fingerprinted sample has outputs on a shared input battery; a pointer-only copy is one whose text Syntology does not serve, because 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, or the licence is outside the permissive list Syntology serves text under (some licences outside it permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on that list). This site shows no code text at all: every sample links to its file in the repository. None of this is a claim about a paper's results, and nothing here is from the archive: this page is Syntology's layer alone.

In numbers Syntology

Syntology ran 127,244 of 294,260 distinct samples harvested for 51,887 papers; 167,016 are unverified. The same code can sit in several papers' repositories, so those samples make 361,576 paper–sample links. One tile per status, counted once per sample:

8,859ran · honoured contract

3,135ran · violated contract

24,451ran · our draft was wrong

9,437ran · fixture could not drive it

81,362ran

167,016unverified

37,273fingerprinted

Licence is a property of each paper's copy, so it is counted per link, not per sample: 124,983 of the 361,576 paper–sample links are pointer only, meaning Syntology does not serve that copy's text (no licence file found, a licence not identified, a permissive licence on a copy whose record is not marked cleared, or a licence outside the permissive list Syntology serves text under). This site shows no code text for any sample; paper pages and name pages link each sample to its file in the repository.

Per sample; the ran count is every status except unverified (the rule each paper page uses), so ran = honoured + violated + our draft was wrong + fixture could not drive it + ran without a contract, and ran + unverified = samples. Read from the graph 2026-09-24. Per-paper detail is on each paper page under "Code Syntology ran".

Every sample on these pages carries its code_sha256, Syntology's identity for that exact code. Agents reach the samples through Syntology's MCP tools: get_harvested_code_for_paper(arxiv_id="…") lists a paper's samples and whether each ran, and get_code(code_sha256="…") fetches one. How to connect.

Find a name

Matches by substring on the exact entry name, in your browser, against the 7,868 names that have a page (those in at least 5 papers). Names in fewer papers have no page and are not in this index; they still appear on their paper pages.

Names by number of papers Syntology

50 shown of 7,868 names with a page, ordered by the number of papers whose harvested code contains a function or class with that exact name. Same-named routines are not asserted to be equivalent; the name is a string, not an identity.

Name	Papers	Papers where it ran	Samples	Unverified	Fingerprinted
conv3x3	1,376	1,241	257	131	1
accuracy	872	422	404	223	55
load_data	700	258	726	493	9
conv1x1	691	619	145	66	0
normalize	638	345	491	274	159
str2bool	533	434	166	72	0
exists	528	513	34	17	7
train	522	131	616	474	4
evaluate	514	175	550	383	8
collate_fn	500	285	384	212	11
load_model	479	59	501	444	1
build_model	469	16	365	349	1
resnet18	432	85	393	328	0
default	430	419	55	14	18
rotate_half	424	409	51	13	38
apply_rotary_pos_emb	393	327	119	48	20
load	392	59	267	210	7
load_config	389	171	367	211	7
get_logger	383	160	287	185	1
test	377	138	411	269	3
window_partition	377	346	113	31	49
MLP	375	335	417	42	116
get_model	368	43	356	316	1
parse_args	355	99	415	315	3
Attention	344	275	428	94	32
load_json	344	171	290	158	4
drop_path	325	283	68	26	17
validate	313	83	306	238	1
gelu	308	265	114	40	70
normalize_answer	304	250	113	49	63
window_reverse	296	268	84	29	0
get_dataset	285	42	292	248	1
preprocess	284	133	258	162	26
Encoder	278	181	380	146	54
count_parameters	278	217	90	52	0
load_checkpoint	277	84	232	174	2
load_dataset	276	94	302	207	0
load_image	273	152	185	124	2
get_pairs	260	243	12	7	0
swish	258	238	41	17	21
Block	252	173	309	98	21
read_json	250	138	173	90	1
repeat_kv	244	233	33	8	21
softmax	240	130	185	89	89
main	235	14	282	268	0
whitespace_clean	230	228	3	2	1
get_optimizer	228	93	171	110	0
Model	227	68	250	173	6
get_data	223	76	248	171	4
split_list	223	212	31	13	16

Which names get a page Syntology

179,672 distinct entry names occur across 294,260 distinct samples. A name gets a page when it appears in at least 5 papers: 7,868 names. The distribution at build time:

Papers per name	Names	Page
1	146,890	no
2	16,565	no
3	5,551	no
4	2,798	no
5–9	4,606	yes
10–19	1,897	yes
20–49	915	yes
50–99	293	yes
100+	157	yes

Names are matched as exact strings, case-sensitively: 194 lowercase forms are shared by two or more names with a page (for example a class and a function that differ only in case), and 213 page addresses needed a numeric suffix to keep such names apart. Behaviour outputs are not in this export; the graph at syntology.ai holds them.