Pith Number

pith:PJW27EJT

pith:2026:PJW27EJTZY2DMHX332FQKTCT7N

not attested not anchored not stored refs pending

Estimating Tail Risks in Language Model Output Distributions

He He, Kathleen McKeown, Raghav Singhal, Rajesh Ranganath, Rico Angell, Zachary Horvitz, Zhou Yu

Creating unsafe versions of language models allows accurate estimation of rare harmful outputs with 10-20 times fewer samples than brute force.

arxiv:2604.22167 v2 · 2026-04-24 · cs.LG · 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{PJW27EJTZY2DMHX332FQKTCT7N}

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

On benchmarks measuring misuse and misalignment, these estimates match brute-force Monte Carlo estimates using 10-20x fewer samples. For example, we can estimate probability of harmful outputs on the order of 10^-4 with just 500 samples.

C2weakest assumption

That unsafe versions of the target model can be constructed such that importance sampling yields unbiased estimates of the original model's harmful output probabilities without introducing systematic bias from the modification process.

C3one line summary

Importance sampling with unsafe model variants estimates tail probabilities of harmful language model outputs using 10-20x fewer samples than brute-force Monte Carlo.

Cited by

1 paper in Pith

Predicting LLM Safety Before Release by Simulating Deployment

Receipt and verification

First computed	2026-06-11T01:09:36.049698Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

7a6daf9133ce34361efbde8b054c53fb699642678948de336cd0f514930a5658

Aliases

arxiv: 2604.22167 · arxiv_version: 2604.22167v2 · doi: 10.48550/arxiv.2604.22167 · pith_short_12: PJW27EJTZY2D · pith_short_16: PJW27EJTZY2DMHX3 · pith_short_8: PJW27EJT

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/PJW27EJTZY2DMHX332FQKTCT7N \
  | 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: 7a6daf9133ce34361efbde8b054c53fb699642678948de336cd0f514930a5658

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "2821c22efcacd59ccb382c8380d4a04172564d3467d64dbac35cdd3708524874",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-04-24T02:30:46Z",
    "title_canon_sha256": "a813d455ed69faccf8b16ac08f27602d5bf9aa1c137fce7e24934da90601bf49"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.22167",
    "kind": "arxiv",
    "version": 2
  }
}