Pith Number

pith:KW54KIDF

pith:2023:KW54KIDFZCTH4FO4O4RS6VPIV7

not attested not anchored not stored refs resolved

LLaMA-Adapter: Efficient Fine-tuning of Language Models with Zero-init Attention

Aojun Zhou, Chris Liu, Hongsheng Li, Jiaming Han, Pan Lu, Peng Gao, Renrui Zhang, Shilin Yan, Xiangfei Hu, Yu Qiao

LLaMA-Adapter adapts frozen LLaMA to follow instructions using only 1.2 million added parameters.

arxiv:2303.16199 v3 · 2023-03-28 · cs.CV · cs.AI · cs.CL · cs.LG · cs.MM

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

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{KW54KIDFZCTH4FO4O4RS6VPIV7}

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

With our efficient training, LLaMA-Adapter can generate high-quality responses, comparable to Alpaca with fully fine-tuned 7B parameters.

C2weakest assumption

The zero-initialized attention mechanism with zero gating adaptively injects the new instructional cues into LLaMA while effectively preserving its pre-trained knowledge.

C3one line summary

LLaMA-Adapter turns frozen LLaMA 7B into a capable instruction follower using only 1.2M new parameters and zero-init attention, matching Alpaca while extending to image-conditioned reasoning on ScienceQA and COCO.

References

278 extracted · 278 resolved · 45 Pith anchors

[1] Alpaca-lora. https://github.com/tloen/alpaca-lora, 2023 2023

[2] Flamingo: a visual language model for few-shot learning 2022

[4] Open llm leaderboard 2023

[5] Language models are few-shot learners 1901

[6] Introduction to the conll-2004 shared task: Semantic role labeling 2004

Cited by

44 papers in Pith

AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration

MM-LIMA: Less Is More for Alignment in Multi-Modal Datasets

What to Say and When to Say it: Live Fitness Coaching as a Testbed for Situated Interaction

LLaVA-Octopus: Unlocking Instruction-Driven Adaptive Projector Fusion for Video Understanding

MathFlow: Enhancing the Perceptual Flow of MLLMs for Visual Mathematical Problems

Receipt and verification

First computed	2026-05-17T23:39:19.917730Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

55bbc52065c8a67e15dc77232f55e8afc43b5a16ab565fd34a0b47466d3502a0

Aliases

arxiv: 2303.16199 · arxiv_version: 2303.16199v3 · doi: 10.48550/arxiv.2303.16199 · pith_short_12: KW54KIDFZCTH · pith_short_16: KW54KIDFZCTH4FO4 · pith_short_8: KW54KIDF

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/KW54KIDFZCTH4FO4O4RS6VPIV7 \
  | 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: 55bbc52065c8a67e15dc77232f55e8afc43b5a16ab565fd34a0b47466d3502a0

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "93d3c5dddc0333e446f23e0119f911c68110fb647f6e9d9a5b4dc35c69d290b1",
    "cross_cats_sorted": [
      "cs.AI",
      "cs.CL",
      "cs.LG",
      "cs.MM"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CV",
    "submitted_at": "2023-03-28T17:59:12Z",
    "title_canon_sha256": "831ffdfc694e4500642984da60782b2e5f1362bd1a2f56a85198952d10fff66a"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2303.16199",
    "kind": "arxiv",
    "version": 3
  }
}