Papers › PaLM: Scaling Language Modeling with Pathways

PaLM: Scaling Language Modeling with Pathways

5 Apr 2022Google Research 2022 4arXiv:2204.02311archive 2025-07-28

Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, Parker Schuh, Kensen Shi, Sasha Tsvyashchenko, Joshua Maynez, Abhishek Rao, Parker Barnes, Yi Tay, Noam Shazeer, Vinodkumar Prabhakaran, Emily Reif, Nan Du, Ben Hutchinson, Reiner Pope, James Bradbury, Jacob Austin, Michael Isard, Guy Gur-Ari, Pengcheng Yin, Toju Duke, Anselm Levskaya, Sanjay Ghemawat, Sunipa Dev, Henryk Michalewski, Xavier Garcia, Vedant Misra, Kevin Robinson, Liam Fedus, Denny Zhou, Daphne Ippolito, David Luan, Hyeontaek Lim, Barret Zoph, Alexander Spiridonov, Ryan Sepassi, David Dohan, Shivani Agrawal, Mark Omernick, Andrew M. Dai, Thanumalayan Sankaranarayana Pillai, Marie Pellat, Aitor Lewkowycz, Erica Moreira, Rewon Child, Oleksandr Polozov, Katherine Lee, Zongwei Zhou, Xuezhi Wang, Brennan Saeta, Mark Diaz, Orhan Firat, Michele Catasta, Jason Wei, Kathy Meier-Hellstern, Douglas Eck, Jeff Dean, Slav Petrov, Noah Fiedel

Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies.

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 37 code samples harvested from 5 repositories linked to this paper: 32 ran (of which 16 constructed an object rather than computing a result; 24 with no instrument failure: 2 honoured, 1 violated, 21 with no contract checked; 8 where Syntology's instrument failed) · 5 unverified. Where Syntology's instrument failed: 6 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: community (archive-listed): 37 samples from 5 repositories, 32 ran (24 with no instrument failure; 8 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.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

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.

chrisociepa/allamo mentioned on GitHubpytorchMITSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Extending Context Window of Large Language Models via Positional Interpolation (arXiv:2306.15595; that paper's own run record: 13 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 2 honoured, 0 violated, 4 with no contract checked; 7 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
foundation-model-stack/fms-fsdp mentioned on GitHubpytorchApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Accelerating Production LLMs with Combined Token/Embedding Speculators (arXiv:2404.19124; that paper's own run record: 5 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 2 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report
google/paxml mentioned on GitHubjaxApache-2.0Syntology: 3 samples, 2 ran (0 instrument-failed) · 1 unverified report
lucidrains/CoCa-pytorch mentioned on GitHubpytorchSyntology: 12 samples, 10 ran (2 instrument-failed) · 2 unverified report
conceptofmind/PaLM-flax jaxSyntology: 9 samples, 8 ran (1 instrument-failed) · 1 unverified report
lucidrains/PaLM-jax jaxSyntology: 6 samples, 5 ran (2 instrument-failed) · 1 unverified report
lucidrains/PaLM-pytorch pytorchSyntology: 7 samples, 7 ran (3 instrument-failed) · 0 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

37 samples harvested: 32 ran (of which 16 constructed an object rather than computing a result; 24 with no instrument failure: 2 honoured, 1 violated, 21 with no contract checked; 8 where Syntology's instrument failed) · 5 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.

2ran · honoured contract
1ran · violated contract
6ran · our draft was wrong
2ran; the contract check could not run
21ran
5unverified

Licence: 0 of the 37 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 5 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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”. For 8 samples that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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.

google/paxml community (archive-listed) 3 samples, 2 ran (0 instrument-failed) · 1 unverified
is_checkpoint_asset paxml/checkpoint_paths.py ran Apache-2.0 (permissive) · c0b06c9deb2f2918 · report fixture: signature x:PosixGPath/0/? · returned bool · level recorded 2026-09-24
metadata_exists paxml/checkpoint_metadata.py ran Apache-2.0 (permissive) · 835cdf8db74062ed · report fixture: signature directory:PosixGPath/0/? · returned bool · level recorded 2026-09-24
restore_metadata paxml/checkpoint_metadata.py unverified Apache-2.0 (permissive) · 73f6618e94cff797 · report
lucidrains/CoCa-pytorch community (archive-listed) 12 samples, 10 ran (2 instrument-failed) · 2 unverified
AllGather in 13 papers coca_pytorch/coca_pytorch.py ran MIT (permissive) · 54e85922a73fb785 · report fixture: returned AllGather · level recorded 2026-09-22
CrossAttention in 85 papers coca_pytorch/coca_pytorch.py ran · metamorphic tier: deterministic MIT (permissive) · 3e79daee4d06c337 · report fixture: returned CrossAttention · level recorded 2026-09-22
EmbedToLatents coca_pytorch/coca_pytorch.py ran · metamorphic tier: deterministic MIT (permissive) · d8d433ddd03a68b1 · report fixture: returned EmbedToLatents · level recorded 2026-09-22
LayerNorm in 121 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
coca_pytorch/coca_pytorch.py ran · metamorphic tier: deterministic MIT (permissive) · cbb7647a3b310ee0 · report fixture: returned LayerNorm · level recorded 2026-09-22
ParallelTransformerBlock coca_pytorch/coca_pytorch.py ran · metamorphic tier: deterministic MIT (permissive) · 7b94868665d25696 · report fixture: returned ParallelTransformerBlock · level recorded 2026-09-22
Residual in 24 papers coca_pytorch/coca_pytorch.py ran · metamorphic tier: deterministic fingerprinted MIT (permissive) · 5fab2dfc0a4ccb06 · report fixture: returned Residual · level recorded 2026-09-22
RotaryEmbedding in 27 papers coca_pytorch/coca_pytorch.py ran · metamorphic tier: well formed MIT (permissive) · 231817c8d194a68b · report fixture: signature max_seq_len:int/0/int, device:str/0/str · returned Tensor · level recorded 2026-09-22
SwiGLU in 9 papers coca_pytorch/coca_pytorch.py ran fingerprinted MIT (permissive) · 84cdb1c0d54b1ef4 · report fixture: signature x:Tensor/3/float · returned Tensor · level recorded 2026-09-24
apply_rotary_pos_emb in 393 papers
census: 14 implementations from 15 papers, 8 distinct outputs across 7 buckets, one shared input per bucket this sample: bucket 3 of 7, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
coca_pytorch/coca_pytorch.py ran · our draft was wrong fingerprinted MIT (permissive) · fd592af960c2230e · report fixture: signature pos:Tensor/2/float, t:Tensor/2/float · returned Tensor · level recorded 2026-09-22
pad_dim_to in 5 papers coca_pytorch/coca_pytorch.py ran · our draft was wrong fingerprinted MIT (permissive) · d3d5bf1b20109dcd · report fixture: signature t:Tensor/3/float, length:int/0/int, dim:int/0/int · returned Tensor · level recorded 2026-09-22
CoCa coca_pytorch/coca_pytorch.py unverified MIT (permissive) · c4b1f7fca53ae6f0 · report
all_gather_variable_batch coca_pytorch/coca_pytorch.py unverified MIT (permissive) · 5079742c167ede96 · report
conceptofmind/PaLM-flax community (archive-listed) 9 samples, 8 ran (1 instrument-failed) · 1 unverified
PaLM palm_flax/palm_flax.py ran · metamorphic tier: deterministic MIT (permissive) · 036d204fef6ec731 · report fixture: returned PaLM · level recorded 2026-09-22
ParallelTransformer palm_flax/palm_flax.py ran · metamorphic tier: deterministic MIT (permissive) · 25a9d6da0d4b31f2 · report fixture: returned ParallelTransformer · level recorded 2026-09-22
ParallelTransformerBlock palm_flax/palm_flax.py ran · metamorphic tier: deterministic MIT (permissive) · 0b005776897ae6b9 · report fixture: returned ParallelTransformerBlock · level recorded 2026-09-22
PreNorm in 34 papers palm_flax/palm_flax.py ran · metamorphic tier: deterministic MIT (permissive) · 4cf5ff951dbba496 · report fixture: returned PreNorm · level recorded 2026-09-22
RotaryEmbedding in 27 papers palm_flax/palm_flax.py ran · metamorphic tier: deterministic MIT (permissive) · c327a17241c2969e · report fixture: returned RotaryEmbedding · level recorded 2026-09-22
apply_rotary_pos_emb in 393 papers
census: 14 implementations from 15 papers, 8 distinct outputs across 7 buckets, one shared input per bucket
palm_flax/palm_flax.py ran · violated contract fingerprinted MIT (permissive) · 37ffe94c2d21b89e · report fixture: signature pos:ArrayImpl/3/float, t:ArrayImpl/3/float · returned ArrayImpl · level recorded 2026-09-22
jax_unstack palm_flax/palm_flax.py ran · our draft was wrong fingerprinted MIT (permissive) · 18a3c5b8f897961c · report fixture: signature x:ArrayImpl/3/float, axis:int/0/int · returned ArrayImpl · level recorded 2026-09-22
rotate_half in 424 papers
census: 24 implementations from 96 papers, 4 distinct outputs across 2 buckets, one shared input per bucket
palm_flax/palm_flax.py ran · honoured contract fingerprinted MIT (permissive) · ca6a06a57596c8b4 · report fixture: signature x:ArrayImpl/2/float · returned ArrayImpl · level recorded 2026-09-22
SwiGLU in 9 papers palm_flax/palm_flax.py unverified MIT (permissive) · f9d416af20c5e200 · report
lucidrains/PaLM-jax community (archive-listed) 6 samples, 5 ran (2 instrument-failed) · 1 unverified
LayerNorm in 121 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
palm_jax/palm.py ran fingerprinted MIT (permissive) · f77c04c0d47070c9 · report fixture: returned LayerNorm · level recorded 2026-09-22
PaLM palm_jax/palm.py ran MIT (permissive) · 75aea79ee837ee93 · report fixture: returned PaLM · level recorded 2026-09-22
apply_rotary_pos_emb in 393 papers
census: 14 implementations from 15 papers, 8 distinct outputs across 7 buckets, one shared input per bucket
palm_jax/palm.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) · 1e2e827e2a6e8c99 · report fixture: signature x:ArrayImpl/2/float, sincos:tuple/1/ArrayImpl · returned ArrayImpl · level recorded 2026-09-22
fixed_pos_embedding in 16 papers palm_jax/palm.py ran; the contract check could not run (RAISES) fingerprinted MIT (permissive) · 78821f282a7ccfa2 · report fixture: signature inv_freq:ArrayImpl/1/float, seq:int/0/int · returned tuple · level recorded 2026-09-22
rotate_every_two in 33 papers
census: 5 implementations from 8 papers, 2 distinct outputs across 2 buckets, one shared input per bucket
palm_jax/palm.py ran · honoured contract fingerprinted MIT (permissive) · 6aa7304ec358cbea · report fixture: signature x:ArrayImpl/2/float · returned ArrayImpl · level recorded 2026-09-22
ParallelTransformerBlock palm_jax/palm.py unverified MIT (permissive) · 51d0ce838f5af6d3 · report
lucidrains/PaLM-pytorch community (archive-listed) 7 samples, 7 ran (3 instrument-failed) · 0 unverified
LayerNorm in 121 papers
census: 5 implementations from 5 papers, 5 distinct outputs across 3 buckets, one shared input per bucket
palm_pytorch/palm_pytorch.py ran MIT (permissive) · c0074f24b1b1dddf · report fixture: returned LayerNorm · level recorded 2026-09-22
PaLM palm_pytorch/palm_pytorch.py ran · our draft was wrong MIT (permissive) · 901f254b57206879 · report fixture: signature dim:int/0/int, num_tokens:int/0/int, depth:int/0/int, dim_head:int/0/int, heads:int/0/int, ff_mult:int/0/int · returned Sequential · level recorded 2026-09-22
ParallelTransformerBlock palm_pytorch/palm_pytorch.py ran MIT (permissive) · 7214c60748cdc2fb · report fixture: returned ParallelTransformerBlock · level recorded 2026-09-22
Residual in 24 papers palm_pytorch/palm_pytorch.py ran fingerprinted MIT (permissive) · b7ffedd1471cf6f7 · report fixture: returned Residual · level recorded 2026-09-22
SwiGLU in 9 papers palm_pytorch/palm_pytorch.py ran fingerprinted MIT (permissive) · 8d2e371b5f3aa503 · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-24
apply_rotary_pos_emb in 393 papers
census: 14 implementations from 15 papers, 8 distinct outputs across 7 buckets, one shared input per bucket this sample: bucket 3 of 7, arg 1: rank 2, kind float, dtype float32 · arg 2: rank 2, kind float, dtype float32
palm_pytorch/palm_pytorch.py ran · our draft was wrong fingerprinted MIT (permissive) · 6d7a8eee9f74827c · report fixture: signature pos:Tensor/2/float, t:Tensor/2/float · returned Tensor · level recorded 2026-09-22
rotate_half in 424 papers
census: 24 implementations from 96 papers, 4 distinct outputs across 2 buckets, one shared input per bucket this sample: bucket 1 of 2, arg 1: rank 2, kind float, dtype float32
palm_pytorch/palm_pytorch.py ran · our draft was wrong fingerprinted MIT (permissive) · 08213c6683993d3b · report fixture: signature x:Tensor/2/float · returned Tensor · level recorded 2026-09-22

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. Siblings are assemblies lifted from this paper in the same batch for the same method: alternative readings kept side by side; none replaces or supersedes another.

4 assemblies on this paper: 4 harvested, 0 generated, listed in 1 groups (siblings lifted in one batch form one group).

4 sibling assemblies for PaLM (harvested), lifted in the same batch on 2026-09-09: alternative readings kept side by side; none replaces or supersedes another.

harvested Assembly for PaLM (graph method: PaLM). 9 bricks; individually on a fixture: 8 ran (of which 5 constructed an object rather than computing a result; 7 with no instrument failure: 1 honoured, 1 violated, 5 with no contract checked; 1 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. PaLM entry · level 1 · ran · 036d204fef6ec731
  2. ParallelTransformer model · level 1 · ran · 25a9d6da0d4b31f2
  3. ParallelTransformerBlock model · level 1 · ran · 0b005776897ae6b9
  4. PreNorm helper · level 1 · ran · 4cf5ff951dbba496
  5. RotaryEmbedding helper · level 1 · ran · c327a17241c2969e
  6. SwiGLU model · level 0 · unverified · f9d416af20c5e200
  7. apply_rotary_pos_emb helper · level 1 · ran · violated contract · 37ffe94c2d21b89e
  8. rotate_half helper · level 1 · ran · honoured contract · ca6a06a57596c8b4
  9. jax_unstack helper · level 1 · ran · our draft was wrong · 18a3c5b8f897961c

harvested Assembly for PaLM (graph method: PaLM). 15 bricks; individually on a fixture: 13 ran (of which 6 constructed an object rather than computing a result; 10 with no instrument failure: 0 honoured, 2 violated, 8 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. CoCa entry · level 0 · unverified · c4b1f7fca53ae6f0
  2. CrossAttention model · level 1 · ran · 3e79daee4d06c337
  3. EmbedToLatents helper · level 1 · ran · d8d433ddd03a68b1
  4. LayerNorm helper · level 1 · ran · cbb7647a3b310ee0
  5. ParallelTransformerBlock model · level 1 · ran · 7b94868665d25696
  6. Residual helper · level 1 · ran · 5fab2dfc0a4ccb06
  7. default helper · level 1 · ran · violated contract · 60fff7c3c400d7ff
  8. exists helper · level 1 · ran · violated contract · aa5486a3650902d8
  9. SwiGLU model · level 1 · ran · 84cdb1c0d54b1ef4
  10. RotaryEmbedding helper · level 1 · ran · 231817c8d194a68b
  11. apply_rotary_pos_emb helper · level 1 · ran · our draft was wrong · fd592af960c2230e
  12. rotate_half helper · level 1 · ran · our draft was wrong · fdd93453f92f9167
  13. AllGather helper · level 1 · ran · 54e85922a73fb785
  14. all_gather_variable_batch helper · level 0 · unverified · 5079742c167ede96
  15. pad_dim_to helper · level 1 · ran · our draft was wrong · d3d5bf1b20109dcd

harvested Assembly for PaLM (graph method: PaLM). 8 bricks; individually on a fixture: 8 ran (of which 3 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 0 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. PaLM entry · level 1 · ran · our draft was wrong · 901f254b57206879
  2. LayerNorm helper · level 1 · ran · c0074f24b1b1dddf
  3. ParallelTransformerBlock model · level 1 · ran · 7214c60748cdc2fb
  4. Residual helper · level 1 · ran · b7ffedd1471cf6f7
  5. RotaryEmbedding helper · level 1 · ran · 231817c8d194a68b
  6. SwiGLU helper · level 1 · ran · 8d2e371b5f3aa503
  7. apply_rotary_pos_emb helper · level 1 · ran · our draft was wrong · 6d7a8eee9f74827c
  8. rotate_half helper · level 1 · ran · our draft was wrong · 08213c6683993d3b

harvested Assembly for PaLM (graph method: PaLM). 6 bricks; individually on a fixture: 5 ran (of which 2 constructed an object rather than computing a result; 3 with no instrument failure: 1 honoured, 0 violated, 2 with no contract checked; 2 where Syntology's instrument failed) · 1 unverified. The assembly has not been run end-to-end.

one of 4 siblings

  1. PaLM entry · level 1 · ran · 75aea79ee837ee93
  2. LayerNorm helper · level 1 · ran · f77c04c0d47070c9
  3. ParallelTransformerBlock model · level 0 · unverified · 51d0ce838f5af6d3
  4. fixed_pos_embedding helper · level 1 · ran; the contract check could not run (RAISES) · 78821f282a7ccfa2
  5. apply_rotary_pos_emb helper · level 1 · ran; the contract check could not run (RAISES) · 1e2e827e2a6e8c99
  6. rotate_every_two helper · level 1 · ran · honoured contract · 6aa7304ec358cbea

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="2204.02311") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Auto DebuggingCode GenerationCommon Sense ReasoningCoreference ResolutionCross-Lingual Question AnsweringFew-Shot LearningHindu KnowledgeKnown UnknownsLanguage ModelingLanguage ModellingLogic Grid PuzzleLogical ReasoningMathMemorizationMulti-task Language UnderstandingMultiple Choice Question Answering (MCQA)Natural Language InferenceNovel ConceptsQuestion AnsweringReading ComprehensionSentence CompletionStrategyQAWinowhy

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Auto Debugging Big-bench Lite PaLM 62B (few-shot, k=5) Exact string match 38.2 #1 of 3 Archive leaderboard report
Auto Debugging Big-bench Lite PaLM 540B (few-shot, k=5) Exact string match 38.2 #2 of 3 Archive leaderboard report
Auto Debugging Big-bench Lite PaLM 8B (few-shot, k=5) Exact string match 14.7 #3 of 3 Archive leaderboard report
Code Generation MBPP PaLM Coder 540B Accuracy 47 #71 of 99 Archive leaderboard report
Code Generation MBPP PaLM 540B Accuracy 36.8 #86 of 99 Archive leaderboard report
Common Sense Reasoning BIG-bench (Known Unknowns) PaLM-540B (few-shot, k=5) Accuracy 73.9 #1 of 3 Archive leaderboard report
Common Sense Reasoning BIG-bench (Winowhy) PaLM-540B (few-shot, k=5) Accuracy 65.9 #1 of 4 Archive leaderboard report
Common Sense Reasoning BIG-bench (Winowhy) PaLM-62B (few-shot, k=5) Accuracy 61.0 #3 of 4 Archive leaderboard report
Common Sense Reasoning ReCoRD PaLM 540B (finetuned) EM 94.0 #4 of 45 Archive leaderboard report
Common Sense Reasoning ReCoRD PaLM 540B (finetuned) F1 94.6 #4 of 45 Archive leaderboard report
Common Sense Reasoning WinoGrande PaLM 540B (0-shot) Accuracy 81.1 #15 of 77 Archive leaderboard report
Common Sense Reasoning WinoGrande PaLM 62B (0-shot) Accuracy 77.0 #21 of 77 Archive leaderboard report
Common Sense Reasoning WinoGrande PaLM-cont 62B (0-shot) Accuracy 77.0 #22 of 77 Archive leaderboard report
Coreference Resolution Winograd Schema Challenge PaLM 540B (fine-tuned) Accuracy 100 #1 of 82 Archive leaderboard report
Coreference Resolution Winograd Schema Challenge PaLM 540B (5-shot) Accuracy 89.5 #11 of 82 Archive leaderboard report
Coreference Resolution Winograd Schema Challenge PaLM 540B (0-shot) Accuracy 89.1 #12 of 82 Archive leaderboard report
Coreference Resolution Winograd Schema Challenge PaLM 540B (1-shot) Accuracy 86.3 #16 of 82 Archive leaderboard report
Cross-Lingual Question Answering TyDiQA-GoldP PaLM-540B (CoT) EM 52.9 #7 of 11 Archive leaderboard report
Extreme Summarization GEM-XSum PaLM (finetuning)-540B Parameters 540 B #2 of 6 Archive leaderboard report
Extreme Summarization GEM-XSum PaLM (finetuning)-540B ROUGE-2 21.2 #2 of 6 Archive leaderboard report
Extreme Summarization GEM-XSum T5-XXL ROUGE-2 21.0 #3 of 6 Archive leaderboard report
Extreme Summarization GEM-XSum PaLM (finetuning)-62B Parameters 62 B #4 of 6 Archive leaderboard report
Extreme Summarization GEM-XSum PaLM (finetuning)-62B ROUGE-2 18.5 #4 of 6 Archive leaderboard report
Language Modelling LAMBADA PaLM-540B (Few-Shot) Accuracy 89.7 #1 of 37 Archive leaderboard report
Language Modelling LAMBADA PaLM-540B (One-Shot) Accuracy 81.8 #9 of 37 Archive leaderboard report
Language Modelling LAMBADA PaLM-540B (Zero-Shot) Accuracy 77.9 #15 of 37 Archive leaderboard report
Logical Reasoning BIG-bench (StrategyQA) PaLM-540B (few-shot, k=5) Accuracy 73.9 #1 of 4 Archive leaderboard report
Logical Reasoning BIG-bench (StrategyQA) PaLM-62B (few-shot, k=5) Accuracy 65.4 #3 of 4 Archive leaderboard report
Memorization BIG-bench (Hindu Knowledge) PaLM-540B (few-shot, k=5) Accuracy 95.4 #1 of 3 Archive leaderboard report
Memorization BIG-bench (Hindu Knowledge) PaLM-62B (few-shot, k=5) Accuracy 77.7 #3 of 3 Archive leaderboard report
Multi-task Language Understanding MGSM PaLM 540B Average (%) 55.0 #6 of 12 Archive leaderboard report
Multiple Choice Question Answering (MCQA) BIG-bench (Novel Concepts) PaLM-540B (few-shot, k=5) Accuracy 71.9 #1 of 4 Archive leaderboard report
Multiple Choice Question Answering (MCQA) BIG-bench (Novel Concepts) PaLM-62B (few-shot, k=5) Accuracy 59.4 #3 of 4 Archive leaderboard report
Natural Language Inference CommitmentBank PaLM 540B (finetuned) Accuracy 100 #1 of 20 Archive leaderboard report
Natural Language Inference CommitmentBank PaLM 540B (finetuned) F1 100 #1 of 20 Archive leaderboard report
Natural Language Inference RTE PaLM 540B (fine-tuned) Accuracy 95.7% #2 of 90 Archive leaderboard report
Natural Language Inference RTE PaLM 540B (5-shot) Accuracy 79.6% #38 of 90 Archive leaderboard report
Natural Language Inference RTE PaLM 540B (1-shot) Accuracy 78.7% #42 of 90 Archive leaderboard report
Natural Language Inference RTE PaLM 540B (0-shot) Accuracy 72.9% #49 of 90 Archive leaderboard report
Question Answering BoolQ PaLM 540B (fine-tuned) Accuracy 92.2 #3 of 65 Archive leaderboard report
Question Answering COPA PaLM 540B (finetuned) Accuracy 100 #1 of 60 Archive leaderboard report
Question Answering MultiRC PaLM 540B (finetuned) EM 69.2 #1 of 30 Archive leaderboard report
Question Answering MultiRC PaLM 540B (finetuned) F1 90.1 #1 of 30 Archive leaderboard report
Question Answering Natural Questions PaLM-540B (Few-Shot, k=64) EM 39.6 #26 of 47 Archive leaderboard report
Question Answering Natural Questions PaLM-540B (One-Shot) EM 29.3 #36 of 47 Archive leaderboard report
Question Answering Natural Questions PaLM-540B (Zero-Shot) EM 21.2 #44 of 47 Archive leaderboard report
Question Answering OBQA PaLM 540B (zero-shot) Accuracy 53.4 #8 of 9 Archive leaderboard report
Question Answering OBQA PaLM 62B (zero-shot) Accuracy 50.4 #9 of 9 Archive leaderboard report
Question Answering TriviaQA PaLM-540B (Few-Shot) EM 81.4 #11 of 56 Archive leaderboard report
Question Answering TriviaQA PaLM-540B (One-Shot) EM 81.4 #12 of 56 Archive leaderboard report
Question Answering TriviaQA PaLM-540B (Zero-Shot) EM 76.9 #17 of 56 Archive leaderboard report
Question Answering WebQuestions PaLM-540B (Few-Shot) EM 43.5 #11 of 37 Archive leaderboard report
Question Answering WebQuestions PaLM-540B (One-Shot) EM 22.6 #30 of 37 Archive leaderboard report
Question Answering WebQuestions PaLM-540B (Zero-Shot) EM 10.6 #34 of 37 Archive leaderboard report
Reading Comprehension RACE PaLM 540B (zero-shot) Accuracy (High) 49.1 #12 of 24 Archive leaderboard report
Reading Comprehension RACE PaLM 540B (zero-shot) Accuracy (Middle) 68.1 #12 of 24 Archive leaderboard report
Reading Comprehension RACE PaLM 62B (zero-shot) Accuracy (High) 47.5 #14 of 24 Archive leaderboard report
Reading Comprehension RACE PaLM 62B (zero-shot) Accuracy (Middle) 64.3 #14 of 24 Archive leaderboard report
Reading Comprehension RACE PaLM 8B (zero-shot) Accuracy (High) 42.3 #19 of 24 Archive leaderboard report
Reading Comprehension RACE PaLM 8B (zero-shot) Accuracy (Middle) 57.9 #19 of 24 Archive leaderboard report
Sentence Completion HellaSwag PaLM-540B (Few-Shot) Accuracy 83.8 #29 of 89 Archive leaderboard report
Sentence Completion HellaSwag PaLM-540B (1-shot) Accuracy 83.6 #30 of 89 Archive leaderboard report
Sentence Completion HellaSwag PaLM-540B (0-shot) Accuracy 83.4 #32 of 89 Archive leaderboard report
Word Sense Disambiguation Words in Context PaLM 540B (finetuned) Accuracy 78.8 #2 of 37 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Introduced by this paper: Parallel Layers

AdafactorAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationMulti-Query AttentionPaLMParallel LayersResidual ConnectionRotary EmbeddingsSentencePieceSoftmaxSwiGLUTransformer

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