Pith Number

pith:5FK7HQTN

pith:2026:5FK7HQTNIH2TDBCPZTCZTHNHAH

not attested not anchored not stored refs resolved

Tracing Persona Vectors Through LLM Pretraining

Dominik Glandorf, Jorge Medina Moreira, Robert West, Tanja K\"aser, Viktor Moskvoretskii

Persona vectors for traits like sycophancy emerge within the first 0.22 percent of LLM pretraining and remain usable for steering the final model.

arxiv:2605.13329 v1 · 2026-05-13 · cs.CL · 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{5FK7HQTNIH2TDBCPZTCZTHNHAH}

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

persona vectors form remarkably early -- within 0.22% of OLMo-3 pretraining -- and remain effective for steering the fully post-trained instruct models. Although core representations are formed early on, persona vectors continue to refine geometrically and semantically throughout pretraining.

C2weakest assumption

That the linear directions identified at early checkpoints represent the same high-level personas as those in the final model and that the elicitation methods isolate these without being confounded by other training dynamics.

C3one line summary

Persona vectors form within the first 0.22% of LLM pretraining and remain effective for steering post-trained models, with continued refinement and transfer to other models.

References

43 extracted · 43 resolved · 5 Pith anchors

[1] System card: Claude Mythos preview 2026

[2] GPT-5.5 system card 2026

[3] Jan Betley, Daniel Chee Hian Tan, Niels Warncke, Anna Sztyber-Betley, Xuchan Bao, Mart´ın Soto, Nathan Labenz, and Owain Evans 2026

[4] Open Problems in Mechanistic Interpretability 2025 · arXiv:2501.16496

[5] A longlist of theories of impact for interpretability

Receipt and verification

First computed	2026-05-18T02:44:48.576627Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

e955f3c26d41f531844fccc5999da701cfadd98e7ef96b2818fb79257165c9fc

Aliases

arxiv: 2605.13329 · arxiv_version: 2605.13329v1 · doi: 10.48550/arxiv.2605.13329 · pith_short_12: 5FK7HQTNIH2T · pith_short_16: 5FK7HQTNIH2TDBCP · pith_short_8: 5FK7HQTN

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/5FK7HQTNIH2TDBCPZTCZTHNHAH \
  | 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: e955f3c26d41f531844fccc5999da701cfadd98e7ef96b2818fb79257165c9fc

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "c7f411b96e19481262bd472de5e00f7cd675e062a5443b4bf5b4fd7f7272b524",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by-sa/4.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-05-13T10:44:23Z",
    "title_canon_sha256": "553b4349d96b595ca66f5501ab112f96fb59c45acf7fcdaa8a02da0a430473a7"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.13329",
    "kind": "arxiv",
    "version": 1
  }
}