Pith Number

pith:YGDLGXJK

pith:2024:YGDLGXJK45MIHSO7LWD27UDDC7

not attested not anchored not stored refs resolved

InternLM2 Technical Report

Aijia Guo, Bin Wang, Chao Xu, Chengqi Lv, Chenya Gu, Chuyu Zhang, Conghui He, Dahua Lin, Demin Song, Fan Wu, Fengzhe Zhou, Fukai Shang, Guoteng Wang, Haijun Lv, Hang Yan, Haochen Ye, Haodong Duan, Haojiong Chen, Hongwei Liu, Huaiyuan Ying, Huanze Tang, Hui Zhao, Jiangning Liu, Jiantao Qiu, Jiaqi Wang, Jiawei Hong, Jiaxing Li, Jiaye Ge, Jia Yu, Jiayu Wang, Jingming Zhuo, Jingwen Li, Jing Yu, Kai Chen, Kai Lv, Kaiwen Liu, Keyu Chen, Kuikun Liu, Li Ma, Linke Ouyang, Linyang Li, Li Zhang, Maosong Cao, Pan Zhang, Pei Chu, Penglong Jiao, Peng Sun, Peng Zhang, Qian Zhao, Qi Fan, Qipeng Guo, Qizhen Weng, Ruijie Zhang, Ruiliang Xu, Rui Wang, Runyuan Ma, Shuaibin Li, Shuo Zhang, Songyang Zhang, Tao Gui, Tao Jiang, Ting Huang, Wei Li, Wenchang Ning, Wenjian Zhang, Wenwei Zhang, Xiaogui Yang, Xiaomeng Zhao, Xiaoran Liu, Xiaoyi Dong, Xin Chen, Xingcheng Zhang, Xingjian Wei, Xinyue Zhang, Xipeng Qiu, Xun Chen, Yang Gao, Yicheng Zou, Yingfan Hu, Yingtong Xiong, Yining Li, Yirong Yan, Yuan Qu, Yudong Wang, Yuhang Zang, Yunfan Shao, Yu Qiao, Yu Sun, Yuzhe Gu, Zaida Zhou, Zehui Chen, Zerun Ma, Zhaoye Fei, Zheng Cai, Zhenjiang Jin, Zhi Chen, Zhihao Sui, Zhikai Lei, Zifan Song, Ziyi Wang

InternLM2 outperforms prior open-source LLMs on 30 benchmarks, long-context tasks up to 200k tokens, and subjective evaluations via staged pre-training and COOL RLHF alignment.

arxiv:2403.17297 v1 · 2024-03-26 · cs.CL · 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{YGDLGXJK45MIHSO7LWD27UDDC7}

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

InternLM2 outperforms its predecessors in comprehensive evaluations across 6 dimensions and 30 benchmarks, long-context modeling, and open-ended subjective evaluations through innovative pre-training and optimization techniques.

C2weakest assumption

That the chosen 30 benchmarks and subjective evaluations fairly measure general capability without hidden selection effects or prompt sensitivity that would change the ranking if different test suites were used.

C3one line summary

InternLM2 is a new open-source LLM that outperforms prior versions on 30 benchmarks and long-context tasks through scaled pre-training to 32k tokens and a conditional online RLHF alignment strategy.

References

172 extracted · 172 resolved · 32 Pith anchors

[1] https://github.com/MicrosoftDocs/azure-docs/blob/main/articles/ai-services/openai/includes/chat-markup-language.md 2024

[2] llama.cpp: Port of facebook's llama model in c/c++. https://github.com/ggerganov/llama.cpp, 2023 2023

[3] GQA: training generalized multi-query transformer models from multi-head checkpoints 2023

[6] Cibench: Evaluating your llms with a code interpreter plugin 2024

[7] Mathbench: Evaluating the theory and application proficiency of llms with a hierarchical mathematics benchmark 2024

Formal links

2 machine-checked theorem links

Cited by

34 papers in Pith

Convergence Without Understanding: When Language Models Agree on Representations but Disagree on Reasoning

Data-Centric Foundation Models in Computational Healthcare: A Survey

From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline

A Scalable Multi-LLM Collaboration System with Retrieval-based Selection and Exploration-Exploitation-Driven Enhancement

Exploring Lightweight Large Language Models for Court View Generation

Receipt and verification

First computed	2026-05-17T23:38:52.642763Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

c186b35d2ae75883c9df5d87afd06317d6869061949e4a452da3c46492fb8e26

Aliases

arxiv: 2403.17297 · arxiv_version: 2403.17297v1 · doi: 10.48550/arxiv.2403.17297 · pith_short_12: YGDLGXJK45MI · pith_short_16: YGDLGXJK45MIHSO7 · pith_short_8: YGDLGXJK

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/YGDLGXJK45MIHSO7LWD27UDDC7 \
  | 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: c186b35d2ae75883c9df5d87afd06317d6869061949e4a452da3c46492fb8e26

Canonical record JSON

{
  "metadata": {
    "abstract_canon_sha256": "9f17aa486b94980a66d7ee8ea79278476f4545a7dfde7563d75d229fe66a619e",
    "cross_cats_sorted": [
      "cs.AI"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "primary_cat": "cs.CL",
    "submitted_at": "2024-03-26T00:53:24Z",
    "title_canon_sha256": "c68c7b4fd42ce43293000a5c5d0e9f23abf3464ed1e8fcaa1cc85474eded6201"
  },
  "schema_version": "1.0",
  "source": {
    "id": "2403.17297",
    "kind": "arxiv",
    "version": 1
  }
}