Fortifying LLM Safety: phi-3’s Responsible AI Alignment

Discover how phi-3 models embody Microsoft’s responsible AI principles, featuring multi-stage safety alignment, rigorous red-teaming, and extensive evaluations across dozens of RAI harm categories for safer LLM deployment.


This content originally appeared on HackerNoon and was authored by Writings, Papers and Blogs on Text Models

Abstract and 1 Introduction

2 Technical Specifications

3 Academic benchmarks

4 Safety

5 Weakness

6 Phi-3-Vision

6.1 Technical Specifications

6.2 Academic benchmarks

6.3 Safety

6.4 Weakness

References

A Example prompt for benchmarks

B Authors (alphabetical)

C Acknowledgements

4 Safety

Phi-3-mini was developed in accordance with Microsoft’s responsible AI principles. The overall approach consisted of safety alignment in post-training, red-teaming, automated testing and evaluations across dozens of RAI harm categories. Helpfulness and harmlessness preference datasets [BJN+ 22, JLD+ 23] with modifications inspired by [BSA+ 24] and multiple in-house generated datasets were leveraged to address the RAI harm categories in safety post-training. An independent red team at Microsoft iteratively examined phi-3-mini to further identify areas of improvement during the post-training process. Based on their feedback, we curated additional datasets tailored to address their insights, thereby refining the post-training dataset. This process resulted in significant decrease of harmful response rates, as shown in Figure 4.

\ Figure 4: Comparison of harmful response percentages by Microsoft AI Red Team between phi-3-mini before and after the safety alignment. Note that the harmful response percentages in this chart are inflated numbers as the red team tried to induce phi-3-mini in an adversarial way to generate harmful responses through multi-turn conversations.

\ Table 1: Comparison of Microsoft internal multi-turn conversation RAI benchmark results of phi-3 models and other models. Note that a lower value indicates a better performance for all metrics in the table.

\ The safety alignment of phi-3-small and phi-3-medium was conducted by undergoing the same red-teaming process, utilizing identical datasets, and incorporating a slightly larger number of samples. Table 1 shows the results of in-house RAI benchmarks [MHJ+ 23] for phi-3 models compared to phi-2 [JBA+ 23], Mistral-7b-v0.1 [JSM+ 23], Gemma 7b [TMH+ 24], and Llama-3-instruct-8b [AI]. This benchmark utilized GPT-4 to simulate multi-turn conversations in five different categories and to evaluate the model responses. Ungroundedness between 0 (fully grounded) and 4 (not grounded) measures if the information in a response is based on a given prompt. In other categories, responses were evaluated in terms of the severity of harmfulness from 0 (no harm) to 7 (extreme harm) and the defect rates (DR-x) were computed as the percentage of samples with the severity score being greater than or equal to x.

\

:::info Authors:

(1) Marah Abdin;

(2) Sam Ade Jacobs;

(3) Ammar Ahmad Awan;

(4) Jyoti Aneja;

(5) Ahmed Awadallah;

(6) Hany Awadalla;

(7) Nguyen Bach;

(8) Amit Bahree;

(9) Arash Bakhtiari;

(10) Jianmin Bao;

(11) Harkirat Behl;

(12) Alon Benhaim;

(13) Misha Bilenko;

(14) Johan Bjorck;

(15) Sébastien Bubeck;

(16) Qin Cai;

(17) Martin Cai;

(18) Caio César Teodoro Mendes;

(19) Weizhu Chen;

(20) Vishrav Chaudhary;

(21) Dong Chen;

(22) Dongdong Chen;

(23) Yen-Chun Chen;

(24) Yi-Ling Chen;

(25) Parul Chopra;

(26) Xiyang Dai;

(27) Allie Del Giorno;

(28) Gustavo de Rosa;

(29) Matthew Dixon;

(30) Ronen Eldan;

(31) Victor Fragoso;

(32) Dan Iter;

(33) Mei Gao; 

(34) Min Gao;

(35) Jianfeng Gao;

(36) Amit Garg;

(37) Abhishek Goswami;

(38) Suriya Gunasekar;

(39) Emman Haider;

(40) Junheng Hao;

(41) Russell J. Hewett;

(42) Jamie Huynh;

(43) Mojan Javaheripi;

(44) Xin Jin;

(45) Piero Kauffmann;

(46) Nikos Karampatziakis;

(47) Dongwoo Kim;

(48) Mahoud Khademi;

(49) Lev Kurilenko; 

(50) James R. Lee;

(51) Yin Tat Lee;

(52) Yuanzhi Li;

(53) Yunsheng Li;

(54) Chen Liang;

(55) Lars Liden;

(56) Ce Liu;

(57) Mengchen Liu;

(58) Weishung Liu;

(59) Eric Lin;

(60) Zeqi Lin;

(61) Chong Luo;

(62) Piyush Madan;

(63) Matt Mazzola;

(64) Arindam Mitra;

(65) Hardik Modi;

(66) Anh Nguyen;

(67) Brandon Norick;

(68) Barun Patra;

(69) Daniel Perez-Becker;

(70) Thomas Portet; 

(71) Reid Pryzant;

(72) Heyang Qin;

(73) Marko Radmilac;

(74) Corby Rosset;

(75) Sambudha Roy; 

(76) Olatunji Ruwase;

(77) Olli Saarikivi;

(78) Amin Saied;

(79) Adil Salim;

(80) Michael Santacroce;

(81) Shital Shah;

(82) Ning Shang;

(83) Hiteshi Sharma;

(84) Swadheen Shukla;

(85) Xia Song;

(86) Masahiro Tanaka;

(87) Andrea Tupini;

(88) Xin Wang;

(89) Lijuan Wang; 

(90) Chunyu Wang;

(91) Yu Wang;

(92) Rachel Ward;

(93) Guanhua Wang;

(94) Philipp Witte; 

(95) Haiping Wu; 

(96) Michael Wyatt; 

(97) Bin Xiao;

(98) Can Xu; 

(99) Jiahang Xu; 

(100) Weijian Xu; 

(101) Sonali Yadav; 

(102) Fan Yang; 

(103) Jianwei Yang;

(104) Ziyi Yang;

(105) Yifan Yang; 

(106) Donghan Yu;

(107) Lu Yuan;

(108) Chengruidong Zhang; 

(109) Cyril Zhang; 

(110) Jianwen Zhang;

(111) Li Lyna Zhang;

(112) Yi Zhang;

(113) Yue Zhang;

(114) Yunan Zhang;

(115) Xiren Zhou.

:::

:::info This paper is available on arxiv under CC BY 4.0 DEED license.

:::

\


This content originally appeared on HackerNoon and was authored by Writings, Papers and Blogs on Text Models


Print Share Comment Cite Upload Translate Updates
APA

Writings, Papers and Blogs on Text Models | Sciencx (2025-07-08T02:00:04+00:00) Fortifying LLM Safety: phi-3’s Responsible AI Alignment. Retrieved from https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/

MLA
" » Fortifying LLM Safety: phi-3’s Responsible AI Alignment." Writings, Papers and Blogs on Text Models | Sciencx - Tuesday July 8, 2025, https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/
HARVARD
Writings, Papers and Blogs on Text Models | Sciencx Tuesday July 8, 2025 » Fortifying LLM Safety: phi-3’s Responsible AI Alignment., viewed ,<https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/>
VANCOUVER
Writings, Papers and Blogs on Text Models | Sciencx - » Fortifying LLM Safety: phi-3’s Responsible AI Alignment. [Internet]. [Accessed ]. Available from: https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/
CHICAGO
" » Fortifying LLM Safety: phi-3’s Responsible AI Alignment." Writings, Papers and Blogs on Text Models | Sciencx - Accessed . https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/
IEEE
" » Fortifying LLM Safety: phi-3’s Responsible AI Alignment." Writings, Papers and Blogs on Text Models | Sciencx [Online]. Available: https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/. [Accessed: ]
rf:citation
» Fortifying LLM Safety: phi-3’s Responsible AI Alignment | Writings, Papers and Blogs on Text Models | Sciencx | https://www.scien.cx/2025/07/08/fortifying-llm-safety-phi-3s-responsible-ai-alignment/ |

Please log in to upload a file.



There are no updates yet.
Click the Upload button above to add an update.

You must be logged in to translate posts. Please log in or register.

Related Posts