Pith Number

pith:O6M274A4

pith:2026:O6M274A44LUGGWHN2K3BP56J2B

not attested not anchored not stored refs resolved

Rethinking Layer Relevance in Large Language Models Beyond Cosine Similarity

Andres Carvallo De Ferari, Christ Devia, Cristian Hinostroza, Denis Parra, Eugenio Herrera-Berg, Jorge F Silva, Rodrigo Toro Icarte

Cosine similarity can be arbitrarily low for a layer that is still essential to an LLM's performance.

arxiv:2605.14075 v1 · 2026-05-13 · cs.LG · cs.CL

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{O6M274A44LUGGWHN2K3BP56J2B}

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

Our theoretical analysis shows that a layer can exhibit an arbitrarily low cosine similarity score while still being crucial to the model's performance. Empirical evidence confirms that the correlation between cosine similarity and actual performance degradation is often weak or moderate.

C2weakest assumption

That removing a single layer and measuring accuracy drop on held-out tasks gives a faithful picture of that layer's contribution inside the intact model, without major compensatory effects from remaining layers.

C3one line summary

Cosine similarity poorly predicts performance degradation from layer removal in LLMs, making direct accuracy-drop ablation a more reliable relevance metric.

References

77 extracted · 77 resolved · 7 Pith anchors

[1] Advances in neural information processing systems , booktitle=NIPS17, year=

[2] All bark and no bite: Rogue dimensions in transformer language models obscure representational quality , author=

[3] A Survey on Transformers in Reinforcement Learning , author=. 2023 , journal= 2023

[4] Breakthroughs in statistics: Methodology and distribution , pages= 1992

[5] Journal of the american statistical association , volume= 1937

Receipt and verification

First computed	2026-05-17T23:39:12.373908Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

7799aff01ce2e86358edd2b617f7c9d078ee5f7225d4c99a86ac22c2ead54a40

Aliases

arxiv: 2605.14075 · arxiv_version: 2605.14075v1 · doi: 10.48550/arxiv.2605.14075 · pith_short_12: O6M274A44LUG · pith_short_16: O6M274A44LUGGWHN · pith_short_8: O6M274A4

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/O6M274A44LUGGWHN2K3BP56J2B \
  | 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: 7799aff01ce2e86358edd2b617f7c9d078ee5f7225d4c99a86ac22c2ead54a40

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "1a2a2a80efea60ded40ab4e08481d98a95ec7b3f2a73a0fbd19be6c9d577819f",
    "cross_cats_sorted": [
      "cs.CL"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-05-13T19:51:25Z",
    "title_canon_sha256": "cce309cae85c94074707337ba9d8d89c0247c9aa05e900492fbcf668743cb6e2"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2605.14075",
    "kind": "arxiv",
    "version": 1
  }
}