Pith Number

pith:XY2G3K54

pith:2026:XY2G3K54XRXGW4UQENJRW7XC2Y

not attested not anchored not stored refs resolved

Benchmarking Mythos-Linked Bug Rediscovery

Arthur Gervais, Isaac David

Even when supplied with the exact target source files, large language models rediscover only six of the intended Mythos-linked systems bugs across 54 attempts.

arxiv:2605.17416 v1 · 2026-05-17 · cs.SE · cs.AI

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{XY2G3K54XRXGW4UQENJRW7XC2Y}

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

Under this favorable target-file scaffold, systems-specific prompting yields only six target matches across 54 counted attempts.

C2weakest assumption

The manual target-matching rubric correctly identifies rediscovery of the intended Mythos bug without bias from the specific language or details omitted from the prompts.

C3one line summary

A benchmarking experiment finds low rediscovery rates for three models on six Mythos-linked bug tasks, with only six target matches across 54 attempts under controlled prompting.

References

55 extracted · 55 resolved · 2 Pith anchors

[1] Partnering with mozilla to improve firefox’s security, 2026 2026

[2] Anthropic. Project glasswing, 2026. URL https://www.anthropic.com/glasswing. Ac- cessed 2026-04-18 2026

[3] Claude mythos preview system card, 2026 2026

[4] Assessing claude mythos preview’s cybersecurity capabilities,

[5] Published 2026-04- 07; accessed 2026-04-18 2026

Receipt and verification

First computed	2026-05-20T00:03:57.341680Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

be346dabbcbc6e6b729023531b7ee2d60b7a933fd4b65260469b948487112420

Aliases

arxiv: 2605.17416 · arxiv_version: 2605.17416v1 · doi: 10.48550/arxiv.2605.17416 · pith_short_12: XY2G3K54XRXG · pith_short_16: XY2G3K54XRXGW4UQ · pith_short_8: XY2G3K54

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/XY2G3K54XRXGW4UQENJRW7XC2Y \
  | 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: be346dabbcbc6e6b729023531b7ee2d60b7a933fd4b65260469b948487112420

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "b34a9ab559bfbe0809b5076fad39b88a247e809cca6c95fde4dbc9bb224f482b",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.SE",
    "submitted_at": "2026-05-17T12:24:11Z",
    "title_canon_sha256": "5b4970652583941957cfafa8246885efc42f035cc19195e08dc4407050e0c755"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.17416",
    "kind": "arxiv",
    "version": 1
  }
}