Pith Number

pith:W3AYUNLZ

pith:2026:W3AYUNLZEN3MUG2TPF2KH65SME

not attested not anchored not stored refs pending

How Alignment Routes: Localizing, Scaling, and Controlling Policy Circuits in Language Models

Gregory N. Frank

Alignment in language models routes safety policies through early attention gates rather than erasing unsafe capabilities.

arxiv:2604.04385 v5 · 2026-04-06 · cs.CL · cs.AI · cs.LG

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{W3AYUNLZEN3MUG2TPF2KH65SME}

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

The safety-trained capability is gated by routing, not removed; modulating the detection-layer signal continuously controls policy from hard refusal through evasion to factual answering, and any encoding that defeats detection-layer pattern matching bypasses the policy regardless of whether deeper layers reconstruct the content.

C2weakest assumption

That interchange interventions and knockout cascades isolate the causal contribution of the identified gate and amplifier heads without substantial side effects on other circuits or on the model's general capability.

C3one line summary

Alignment policy in language models is implemented as an early-commitment routing circuit of detection gates and amplifier heads that can be localized, scaled, and directly controlled without removing the underlying capability.

Cited by

2 papers in Pith

Fast Multi-dimensional Refusal Subspaces via RFM-AGOP

Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications

Receipt and verification

First computed	2026-06-30T02:17:20.138356Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

b6c18a35792376ca1b537974a3fbb2611876e15fd7f7ed7276e3dde75a33cdf6

Aliases

arxiv: 2604.04385 · arxiv_version: 2604.04385v5 · doi: 10.48550/arxiv.2604.04385 · pith_short_12: W3AYUNLZEN3M · pith_short_16: W3AYUNLZEN3MUG2T · pith_short_8: W3AYUNLZ

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/W3AYUNLZEN3MUG2TPF2KH65SME \
  | 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: b6c18a35792376ca1b537974a3fbb2611876e15fd7f7ed7276e3dde75a33cdf6

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "f522f2b99c1fb0ffb7e53ab70c8fefe8c3473976e4c4ffa1703bcdf5cbd639ec",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.LG"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2026-04-06T03:20:37Z",
    "title_canon_sha256": "2535a2bd1b60beef235b319f09b2c08e0dc47b227dca504ee3bdfb81f47b9caf"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2604.04385",
    "kind": "arxiv",
    "version": 5
  }
}