Pith Number

pith:L4DYAMP4

pith:2020:L4DYAMP4CO2UVHDPGSULJ2LCZG

not attested not anchored not stored refs pending

Big Bird: Transformers for Longer Sequences

Amr Ahmed, Anirudh Ravula, Avinava Dubey, Chris Alberti, Guru Guruganesh, Joshua Ainslie, Li Yang, Manzil Zaheer, Philip Pham, Qifan Wang, Santiago Ontanon

BigBird's sparse attention preserves universal approximation and Turing completeness while scaling transformers to much longer sequences.

arxiv:2007.14062 v2 · 2020-07-28 · cs.LG · cs.CL · stat.ML

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{L4DYAMP4CO2UVHDPGSULJ2LCZG}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest claim

We show that BigBird is a universal approximator of sequence functions and is Turing complete, thereby preserving these properties of the quadratic, full attention model.

C2weakest assumption

The chosen combination of global, local, and random attention tokens is sufficient to retain the expressive power of full attention for the tasks considered.

C3one line summary

BigBird replaces full attention in Transformers with a sparse pattern that achieves linear complexity while remaining a universal approximator and Turing complete.

Formal links

1 machine-checked theorem link

Cited by

20 papers in Pith

Dynamic Video Generation: Shaping Video Generation Across Time and Space

Block-Based Double Decoders

Measuring Investor Learning in Private Markets: A Sequential LLM-Bayesian Analysis of Expert Network Calls

LPC-SM: Local Predictive Coding and Sparse Memory for Long-Context Language Modeling

Gated Linear Attention Transformers with Hardware-Efficient Training

Receipt and verification

First computed	2026-05-17T23:38:45.913475Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

5f078031fc13b54a9c6f34a8b4e962c9ba684e49e63756b2309e9e3feca98726

Aliases

arxiv: 2007.14062 · arxiv_version: 2007.14062v2 · doi: 10.48550/arxiv.2007.14062 · pith_short_12: L4DYAMP4CO2U · pith_short_16: L4DYAMP4CO2UVHDP · pith_short_8: L4DYAMP4

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/L4DYAMP4CO2UVHDPGSULJ2LCZG \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: 5f078031fc13b54a9c6f34a8b4e962c9ba684e49e63756b2309e9e3feca98726

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "715aa306ef7fc43efc29c6b8f31619884065434b7f7310d8bce8aba1b16c8446",
    "cross_cats_sorted": [
      "cs.CL",
      "stat.ML"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2020-07-28T08:34:04Z",
    "title_canon_sha256": "fe93aca2c56e0c722572fa55194f449de0bdf4351c41094903d025e445c17514"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2007.14062",
    "kind": "arxiv",
    "version": 2
  }
}