Pith Number

pith:TDQ2LKW7

pith:2026:TDQ2LKW76ZLL5LCBPDNWIT2VQL

not attested not anchored not stored refs resolved

Collaborative Parameter Learning: Mitigating Forgetting via Parameter-Level Gradient Analysis

Haolin Li, Jiandong Gao, Ji Wu, Kaili Zheng, Kaiwen Wang, Mutian Yang, Qi Wang, Yuguang Wang, Yutong Chen, Zisen Zhan

Collaborative Parameter Learning freezes conflicting parameters to let large language models acquire new knowledge while retaining old capabilities.

arxiv:2601.21577 v2 · 2026-01-29 · 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{TDQ2LKW76ZLL5LCBPDNWIT2VQL}

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

Experiments comparing CPL with seven baseline methods show that CPL learns 20.2% to 48.2% more questions with negligible forgetting, while reducing peak VRAM by approximately 3 GB per billion model parameters and computation time by 16.5 percent.

C2weakest assumption

That the parameter-wise gradient contributions observed during a single training run reliably classify parameters as conflicting or collaborative in a way that generalizes across models, tasks, and future updates without requiring per-run reclassification.

C3one line summary

Collaborative Parameter Learning freezes 50-75% of parameters whose updates cause forgetting and updates only the 25-50% that mitigate it, allowing LLMs to learn 20-48% more new questions with negligible forgetting and lower compute cost.

References

23 extracted · 23 resolved · 8 Pith anchors

[1] GPT-4 Technical Report · arXiv:2303.08774

[2] Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge · arXiv:1803.05457

[3] Time sensitive knowledge editing through efficient finetuning.arXiv preprint arXiv:2406.04496,

[4] The Llama 3 Herd of Models · arXiv:2407.21783

[5] Adams, Jens-Michalis Papaioannou, Paul Grundmann, Tom Oberhauser, Alexei Figueroa, Alexander Löser, Daniel Truhn, and Keno K

Receipt and verification

First computed	2026-05-18T03:09:24.160551Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

98e1a5aadff656beac4178db644f5582ce0e51093433fce4ce9a7b3c052298b7

Aliases

arxiv: 2601.21577 · arxiv_version: 2601.21577v2 · doi: 10.48550/arxiv.2601.21577 · pith_short_12: TDQ2LKW76ZLL · pith_short_16: TDQ2LKW76ZLL5LCB · pith_short_8: TDQ2LKW7

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/TDQ2LKW76ZLL5LCBPDNWIT2VQL \
  | 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: 98e1a5aadff656beac4178db644f5582ce0e51093433fce4ce9a7b3c052298b7

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "3923ef78e3e49fb8394c59c6468f84c2f3483802e13541496ae4597dc89cb34f",
    "cross_cats_sorted": [],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.LG",
    "submitted_at": "2026-01-29T11:42:30Z",
    "title_canon_sha256": "6c07a804fc51b781711d37be73fe012ecc0de7c4d77b1cf4d3ac38569015916c"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2601.21577",
    "kind": "arxiv",
    "version": 2
  }
}