Pith Number

pith:FVMKECSP

pith:2023:FVMKECSPVYLV34B5JLGYBO62JC

not attested not anchored not stored refs pending

Inference-Time Intervention: Eliciting Truthful Answers from a Language Model

Fernanda Vi\'egas, Hanspeter Pfister, Kenneth Li, Martin Wattenberg, Oam Patel

Shifting activations in a few attention heads during inference raises LLM truthfulness on TruthfulQA from 32.5 percent to 65.1 percent.

arxiv:2306.03341 v6 · 2023-06-06 · cs.LG · cs.AI · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{FVMKECSPVYLV34B5JLGYBO62JC}

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

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

ITI improves truthfulness of an instruction-finetuned LLaMA (Alpaca) on TruthfulQA from 32.5% to 65.1% by shifting activations along learned directions in a limited number of attention heads.

C2weakest assumption

That the truthful directions identified from a few hundred examples remain effective and stable across unseen prompts and do not introduce systematic new errors beyond the documented truthfulness-helpfulness tradeoff.

C3one line summary

ITI shifts activations in limited attention heads using directions found from a few hundred examples, raising Alpaca's TruthfulQA truthfulness from 32.5% to 65.1% while allowing tunable tradeoff with helpfulness.

Formal links

3 machine-checked theorem links

Cited by

22 papers in Pith

Chain-of-Verification Reduces Hallucination in Large Language Models

DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models

A Comprehensive Survey of Hallucination Mitigation Techniques in Large Language Models

Correcting Influence: Unboxing LLM Outputs with Orthogonal Latent Spaces

Not Just RLHF: Why Alignment Alone Won't Fix Multi-Agent Sycophancy

Receipt and verification

First computed	2026-05-17T23:38:52.912988Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

2d58a20a4fae175df03d4acd80bbda48bf45972da31a816513c1ca8ae6f201c5

Aliases

arxiv: 2306.03341 · arxiv_version: 2306.03341v6 · doi: 10.48550/arxiv.2306.03341 · pith_short_12: FVMKECSPVYLV · pith_short_16: FVMKECSPVYLV34B5 · pith_short_8: FVMKECSP

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/FVMKECSPVYLV34B5JLGYBO62JC \
  | 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: 2d58a20a4fae175df03d4acd80bbda48bf45972da31a816513c1ca8ae6f201c5

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "56ab04632bca05f6f1938d141d06fa754445e1e27475c0d20eb27848e8ba8740",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2023-06-06T01:26:53Z",
    "title_canon_sha256": "807aff8083d26511bae73c8d06f1bdd9b9330faec41de63b95bac2266249f21b"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2306.03341",
    "kind": "arxiv",
    "version": 6
  }
}