Home › Census › top_k_top_p_filtering

top_k_top_p_filtering

Syntologyfunction-name censuscensus 2026-09-22battery b986f7e04d79all samples with this name

top_k_top_p_filtering: 8 implementations from 8 papers ran on one shared input (census 2026-09-22, battery b986f7e04d79); they produced 4 distinct outputs across 2 buckets, one shared input per bucket.

Identical values to six decimals (the recorded digest) on the shared input are agreement on those inputs, not equivalence. Implementations are compared only within one bucket, the positional (rank, kind, dtype) of each array argument; the argument name is not part of the key because the harness draws the shared array from (rank, kind) and casts it to the dtype, whatever the name; each bucket's shared input is fixed by that key, so members of one bucket saw bitwise-identical inputs under their own scalar arguments. A cluster is the set of members whose recorded output digest is identical. Nothing here says which computation a paper's method intended, and nothing reproduces a paper's results. Each member's behaviour fingerprint (its outputs on the shared battery) is not in this export and is not shown here; get_code(code_sha256="…") returns this sample's fingerprint on the shared battery, with the battery's shuffled null and lift; not shown on this site (how to connect); click a member's code_sha256 to copy that call.

Not compared, and not in the tables or the counts above:

Bucket 1 of 2: arg 1: rank 2, kind float, dtype float32

6 implementations from 6 papers share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 2 distinct outputs, largest cluster first. Values are the first 8 of the recorded output, flattened.

Cluster (same digest to six decimals)MembersShared output on this bucket's input
5 implementations
5 papers
e84b62b578f6
recorded values identical

symbolicregression/model/transformer.py 02e71d97ed55e54d
recorded logits:2/float/float32; top_k=2, top_p=0.9, filter_value=-inf, min_tokens_to_keep=1

modeling.py 541fe035a87b91ca
recorded logits:2/float/float32; top_k=2, top_p=0.9, filter_value=-inf

poet/models/poet.py 733f20ae4d66d5ae
recorded logits:2/float/float32; top_k=2, top_p=0.9, filter_value=-inf, min_tokens_to_keep=1

generation/dexperts_generation.py 8c347cb668f95910
recorded logits:2/float/float32; top_k=2, top_p=0.9, filter_value=-inf, min_tokens_to_keep=1

proxy_model/dexpert.py f24ceca3d1604ba8
recorded logits:2/float/float32; top_k=2, top_p=0.9, filter_value=-inf, min_tokens_to_keep=1
[1.80263, -inf, -inf, -inf, -inf, -inf, -inf, 0.572889, …]
shape [4, 8] · float32 · Tensor · non-finite
1 implementation
1 paper
a9b629c9f291

symbolicregression/model/transformer.py f434565e2011ab22
recorded logits:2/float/float32; top_k=3, top_p=0.9, filter_value=-inf, min_tokens_to_keep=1
[1.80263, -inf, 0.476053, -inf, -inf, -inf, -inf, 0.572889, …]
shape [4, 8] · float32 · Tensor · non-finite

Bucket 2 of 2: arg 1: rank 1, kind float, dtype float32

2 implementations from 2 papers share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 2 distinct outputs, largest cluster first. Values are the first 8 of the recorded output, flattened.

Cluster (same digest to six decimals)MembersShared output on this bucket's input
1 implementation
1 paper
0922c676923f

seq_queries/sample.py d111355f8bf308f3
recorded logits:1/float/float32; top_k=5, top_p=0.9, min_var=False, filter_value=-inf, is_log_prob=False
[-inf, 0.157131, 0.897612, -inf, -0.176256, 0.0530317, -inf, -0.177533]
shape [8] · float32 · Tensor · non-finite
1 implementation
1 paper
a7813f1ae6c9

main_DBS.py a2d595fc39ea3025
recorded logits:1/float/float32; top_k=3, top_p=0.9, filter_value=-inf
[-inf, 0.157131, 0.897612, -inf, -inf, 0.0530317, -inf, -inf]
shape [8] · float32 · Tensor · non-finite

Identical values to six decimals (the recorded digest) on the shared input are agreement on those inputs, not equivalence; where a cluster's members carry recorded values, the largest difference among them is shown under the cluster. Paper titles are the archive's archive 2025-07-28 where the paper is in the archive and the graph's where it was added by Syntology; papers with no page here are shown by their recorded paper id only. A paper count above the implementation count means one implementation (one code sha) is held from several papers' repositories and counts once. Per-sample status, licence and fingerprint records for each paper are on its paper page. JSON twin: /census/top-k-top-p-filtering.json.

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