Pith Number

pith:X5R6XHVJ

pith:2026:X5R6XHVJKOLO56LFPET4YJ5P5A

not attested not anchored not stored refs resolved

EndPrompt: Efficient Long-Context Extension via Terminal Anchoring

Dawei Yin, Fang Wang, Han Tian, Haoyi Xiong, Jiamin Chen, Jiashu Zhao, Jinman Zhao, Luxuan Chen, Rui Kong, Shuaiqiang Wang, Xinran Chen, Yuchen Li

EndPrompt extends LLM context windows to 64K by training only on short sequences with a terminal prompt anchored at target positions.

arxiv:2605.14589 v1 · 2026-05-14 · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{X5R6XHVJKOLO56LFPET4YJ5P5A}

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

long-context generalization can be induced from sparse positional supervision, challenging the prevailing assumption that dense long-sequence training is necessary for reliable context-window extension.

C2weakest assumption

That assigning positional indices near the target length to a brief terminal prompt in short sequences preserves the necessary relative distances and semantic continuity for effective long-context learning without introducing artifacts from the artificial split.

C3one line summary

EndPrompt induces reliable long-context generalization in LLaMA models from sparse positional supervision via a two-segment short-sequence construction with terminal anchoring.

References

35 extracted · 35 resolved · 10 Pith anchors

[1] Longalign: A recipe for long context alignment of large language models 2024

[2] LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding 2023 · arXiv:2308.14508

[3] Longformer: The Long-Document Transformer 2004 · arXiv:2004.05150

[4] Lexglue: A benchmark dataset for legal language understanding in english.Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics, 2022 2022

[5] L-Eval: Instituting standardized evaluation for long context language models 2023

Formal links

2 machine-checked theorem links

Receipt and verification

First computed	2026-05-17T23:39:05.268119Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

bf63eb9ea95396eef9657927cc27afe8040909cfdacdb83c1c8df3399b9b9c5b

Aliases

arxiv: 2605.14589 · arxiv_version: 2605.14589v1 · doi: 10.48550/arxiv.2605.14589 · pith_short_12: X5R6XHVJKOLO · pith_short_16: X5R6XHVJKOLO56LF · pith_short_8: X5R6XHVJ

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/X5R6XHVJKOLO56LFPET4YJ5P5A \
  | 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: bf63eb9ea95396eef9657927cc27afe8040909cfdacdb83c1c8df3399b9b9c5b

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "49a76783063716cabb3487ccd241489241ded9027beb749d517950142bd96167",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-14T09:00:03Z",
    "title_canon_sha256": "ed77087556ed0d48317c2d5798de214de2707e733277664f483df5c8217b5ab2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14589",
    "kind": "arxiv",
    "version": 1
  }
}