Home › Census › masked_softmax
masked_softmax
masked_softmax: 10 implementations from 10 papers ran on one shared input (census 2026-09-22, battery b986f7e04d79); they produced 10 distinct outputs across 6 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.
Not compared, and not in the tables or the counts above:
- 3 more implementations of this name were examined but did not run on the shared input (RuntimeError 2, IndexError 1).
- No implementation of this name ran on its own fixture arguments only: every one that ran took an array argument.
- Every compared output was digested.
Bucket 1 of 6: arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
5 implementations from 5 papers share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 5 distinct outputs, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper 564451c2ada1 |
Model_code/models.py 590b22f9 recorded vec:2/float/float32, mask:2/float/float32; dim=1 |
[1.89082, -0.656062, 0.0242994, -0.0453132, -0.0830124, -0.12126, -0.0418446, 0.0323687, …] shape [4, 8] · float32 · Tensor |
| 1 implementation 1 paper cc364006c12c |
excord.py 971291e2 recorded vector:2/float/float32, mask:2/float/float32; dim=-1, memory_efficient=False, mask_fill_value=-1e+32 |
[1.89082, -0.656061, 0.0242994, -0.0453132, -0.0830124, -0.12126, -0.0418446, 0.0323687, …] shape [4, 8] · float32 · Tensor |
| 1 implementation 1 paper dec3904e83f8 |
model/seq2seqPTR/src/ptrbert.py 42fb56de recorded vector:2/float/float32, mask:2/float/float32; dim=-1, mask_fill_value=-1e+32 |
[0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125, 0.125, …] shape [4, 8] · float32 · Tensor |
| 1 implementation 1 paper f2360ef3a23d |
alfworld/agents/modules/model.py e5002de1 recorded x:2/float/float32, m:2/float/float32; axis=-1 |
[1.89082, -0.656061, 0.0242994, -0.0453131, -0.0830123, -0.12126, -0.0418446, 0.0323686, …] shape [4, 8] · float32 · Tensor |
| 1 implementation 1 paper feda954aa726 |
model/basic.py c8b051f5 recorded logits:2/float/float32, mask:2/float/float32; dim=-1 |
[0.99728, -0.030178, 0.0698943, -0.0315789, -0.033385, -0.0335388, -0.0311575, 0.0926643, …] shape [4, 8] · float32 · Tensor |
Bucket 2 of 6: arg 1: rank 1, kind float, dtype float64 · arg 2: rank 1, kind int, dtype int64
1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper 6724e8d7aae6 |
private/grammar.py fd360cf9 recorded logit:1/float/float64, mask:1/int/int64 |
[0.0237781, 0.0343512, 0.216098, 0, 0.0246124, 0.0619102, 0.0115215, 0.049162] shape [8] · float64 · ndarray |
Bucket 3 of 6: arg 1: rank 2, kind float, dtype float32 · arg 2: rank 1, kind int, dtype int64
1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper 7ebb8c96de1f |
text_rationale_attention/attention_model.py 79d2a064 recorded attn_odds:2/float/float32, lens:1/int/int64 |
[0.965657, 0.0343432, 0, 0, 0, 0, 0, 0, …] shape [4, 8] · float32 · Tensor · non-finite |
Bucket 4 of 6: arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind bool, dtype bool
1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper 728c40cf2066 |
modules/named_models.py 27a75066 recorded vec:2/float/float32, mask:2/bool/bool; dim=-1, epsilon=0.001 |
[0.735413, 0.0261547, 0.195167, 0, 0, 0.0432645, 0, 0, …] shape [4, 8] · float32 · Tensor |
Bucket 5 of 6: arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind int, dtype int64
1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper 053f560feb52 |
hxetda/hxetda.py d2b7df2f recorded vec:2/float/float32, mask:2/int/int64; dim=1, epsilon=1e-10 |
[-0.901006, 3.06792, -0.412553, 0.00614049, 0.00496763, 2.06261, -0.662761, 0.0217297, …] shape [4, 8] · float32 · Tensor |
Bucket 6 of 6: arg 1: rank 3, kind float, dtype float32 · arg 2: rank 3, kind bool, dtype bool
1 implementation from 1 paper share this bucket (rank, kind, dtype of each array argument, positional; each member's recorded signature, argument name included, is shown under it); 1 distinct output, largest cluster first. Values are the first 8 of the recorded output, flattened.
| Cluster (same digest to six decimals) | Members | Shared output on this bucket's input |
|---|---|---|
| 1 implementation 1 paper d189fa9dc97b |
model/modules/Attention.py fbe8aa7e recorded attn_odds:3/float/float32, masks:3/bool/bool |
[0, 0.813472, 0.0429335, 0, 0, 0.0360065, 0.107588, 0, …] shape [2, 4, 8] · float32 · Tensor |
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/masked-softmax.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