Pith Number

pith:BWA3S2Y3

pith:2026:BWA3S2Y3ITO6EYF2RZLBDD6I4U

not attested not anchored not stored refs resolved

Multi-Turn Neural Transparency: Surfacing Neural Activations Improves User Calibration to LLM Behavioral Drift

Anthony Baez, Pat Pataranutaporn, Sheer Karny

Surfacing an LLM's internal neural activations in real time helps users better anticipate and evaluate shifts in chatbot behavior across a conversation.

arxiv:2605.15455 v1 · 2026-05-14 · cs.HC

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{BWA3S2Y3ITO6EYF2RZLBDD6I4U}

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

Participants without visualization struggled to accurately evaluate traits (RMSE ≈ 0.6-0.7), while the inclusion of neural transparency significantly improved both anticipation and evaluation compared to no visualization (d = -0.34 to -0.49). The multi-turn dynamic visualization additionally outperformed the static single-turn visualization on holistic evaluation of model behavior (d = -0.32). Transparency also reduced overconfidence.

C2weakest assumption

The behavioral vectors identified via contrastive system prompts accurately and stably represent the expression of the six personality traits in the LLM's activation space across different models and contexts.

C3one line summary

Multi-turn neural transparency using behavioral vectors and dynamic visualizations improves user anticipation and evaluation of LLM trait expression while reducing overconfidence, per a randomized study with 246 participants.

References

52 extracted · 52 resolved · 6 Pith anchors

[1] Understanding intermediate layers using linear classifier probes 2016 · arXiv:1610.01644

[2] Emmanuel Ameisen, Jack Lindsey, Adam Pearce, Wes Gurnee, Nicholas L. Turner, Brian Chen, Craig Citro, David Abrahams, Shan Carter, Basil Hosmer, Jonathan Marcus, Michael Sklar, Adly Templeton, Trenton 2025

[3] Chayapatr Archiwaranguprok, Constanze Albrecht, Pattie Maes, Karrie Kara- halios, and Pat Pataranutaporn. 2025. Simulating Psychological Risks in Human- AI Interactions: Real-Case Informed Modeling of 2025

[4] Andy Arditi, Oscar Obeso, Aaquib Syed, Daniel Paleka, Nina Panickssery, Wes Gurnee, and Neel Nanda. 2024. Refusal in language models is mediated by a single direction.Advances in Neural Information Pr 2024

[5] Persona Vectors: Monitoring and Controlling Character Traits in Language Models 2025 · arXiv:2507.21509

Formal links

1 machine-checked theorem link

Receipt and verification

First computed	2026-05-20T00:00:59.478264Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

0d81b96b1b44dde260ba8e56118fc8e519a543a5606796f9d4a8229d90d9af53

Aliases

arxiv: 2605.15455 · arxiv_version: 2605.15455v1 · doi: 10.48550/arxiv.2605.15455 · pith_short_12: BWA3S2Y3ITO6 · pith_short_16: BWA3S2Y3ITO6EYF2 · pith_short_8: BWA3S2Y3

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/BWA3S2Y3ITO6EYF2RZLBDD6I4U \
  | 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: 0d81b96b1b44dde260ba8e56118fc8e519a543a5606796f9d4a8229d90d9af53

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "cef7bff4139af60dbcbf57fe5378fa7d7eba972823e1f5c43915982cc740dddf",
    "cross_cats_sorted": [],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.HC",
    "submitted_at": "2026-05-14T22:37:56Z",
    "title_canon_sha256": "4e94b40a85ef2f6b8f2dff3a0f4fe9b1e5722ded430fba11594ca3e9ef6a6ac2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.15455",
    "kind": "arxiv",
    "version": 1
  }
}