04 / Impact & talks

Impact & talks

Recognition

Awards & recognition

Personal honors

Award

Beijing Science and Technology Progress Award, Second Prize

The project “Key Technologies and Applications of Large-Model Safety” received the 2025 Beijing Science and Technology Progress Award, Second Prize; Lin Sun participated as the second project contributor.

Talent and industry recognition

Chaoyang Phoenix Talent

Selected for the Chaoyang Phoenix Talent program.

Talent recognition

Capital Cybersecurity Talent

Awarded the title of Capital Cybersecurity Talent.

Team achievements & recognition

Appraisal

CAAI technology achievement appraisal

“Key Technologies and Applications of Large-Model Safety” passed the technology achievement appraisal of the Chinese Association for Artificial Intelligence. The overall technology reached an internationally advanced level, while the detection–response separation technology reached an internationally leading level.

Open research

TinyR1 open model and paper

TinyR1 achieved mathematical performance close to DeepSeek-R1 at about 5% of its parameter scale, received coverage from the National Natural Science Foundation of China, and reached the Hugging Face Trending Top 5.

Team evaluation

Two-time SafetyAI Bench first place

The 360 LLM safety team ranked first in two consecutive quarterly SafetyAI Bench evaluations.

Team evaluation

SuperCLUE-RAG first place, November 2024

The 360 Zhibrain RAG team solution ranked first on the November 2024 SuperCLUE-RAG leaderboard.

Product case

360 Zh Dun selected as a national LLM safety product

360 Zh Dun, a company LLM safety product, was selected as a public case.

Team honor

National Worker Pioneer

360 Zhibrain Division received the National Worker Pioneer honor.

Team competition

Fifth China Artificial Intelligence Competition · Grade A certificate

Received a Grade A certificate in the large-model safety attack-and-defense track of the Fifth China Artificial Intelligence Competition.

Competition award

Xiong’an Cybersecurity Technology Application Competition · Second Prize

Received Second Prize in the Xiong’an Cybersecurity Technology Application Competition.

Product evaluation

LLM safety guardrail evaluation · First place

During China Cybersecurity Week 2026, 360 Large-Model Guardian ranked first in the LLM safety guardrail scenario of application testing organized by CNCERT/CC.

Patents

Patents

LLM safety & model optimization5

  1. Large-model safety protection method, apparatus, device, and storage medium
  2. Large-model optimization method, apparatus, device, and storage medium
  3. Risk detection method, apparatus, device, storage medium, and product
  4. Model optimization method, apparatus, device, storage medium, and product
  5. Answer generation method, apparatus, device, storage medium, and product

Knowledge bases & corpora3

  1. Corpus construction method, apparatus, device, storage medium, and product
  2. Knowledge-base construction method, apparatus, device, storage medium, and product
  3. Data recall method and system, interaction method and system

Search, QA & interaction8

  1. Method and apparatus for obtaining thematic QA data through search
  2. Method and apparatus for presenting QA search results based on search terms
  3. Search-based question publication method and apparatus
  4. Search-result integration method and apparatus
  5. QA-based search-result adjustment method and apparatus
  6. Method and apparatus for providing QA-pair search results
  7. Apparatus and method for determining crawling frequency of web resources
  8. Apparatus and method for determining the relevance of QA pairs

Intelligent interaction & multimedia2

  1. Script-based video presentation method, apparatus, and storage medium
  2. Intelligent interaction method and apparatus based on natural-language syntax analysis
Standards & reports

Standards & reports

Published national standard

Artificial intelligence — Large models, Part 2: Evaluation indicators and methods

GB/T 45288.2—2025
Published national standard

Artificial intelligence — Knowledge graph technology framework

GB/T 42131-2022
Published industry standard

Competency requirements for virtual-reality professionals

SJ/T 11914-2023
Technical report

Report on Large Language Model Safety (2024)

Talks

Public talks

LLM safety governance at 360COSCon’25 · Jingzhi community
Listed speaker · 2025-12-07
LLM safety practice at 360CICC workshop on large models and decision intelligence
Listed presenter · 2025-05-24
Exploring solutions to LLM hallucinationsDataFunSummit 2024 · RAG forum
Listed guest speaker · 2024-08-03