Runjin Chen (@RunjinChen) profile picture

Runjin Chen

@RunjinChen

Member of Technical Staff @AnthropicAI | PH.D. student @UTAustin @VITAGroupUT | Previously BS/MS @sjtu1896

Joined August 2023
  • Tweets15
  • Following55
  • Followers561
1Media
Photo by Runjin Chen (@RunjinChen)
New Anthropic Research: Persona Vectors We can: 1. Monitor how a model’s personality is changing during a conversation, or over training 2. Mitigate undesirable persona shifts during development or prevent during training. 3. Identify training data that leads to shift
Anthropic (@AnthropicAI) profile pictureAnthropic
@AnthropicAI1 Aug 2025
New Anthropic research: Persona vectors. Language models sometimes go haywire and slip into weird and unsettling personas. Why? In a new paper, we find “persona vectors"—neural activity patterns controlling traits like evil, sycophancy, or hallucination.
Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

ALT Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

8
22
2
223
20,533
Runjin Chen retweeted
Jack Lindsey (@Jack_W_Lindsey) profile picture
Jack Lindsey@Jack_W_LindseyApr 7
Before limited-releasing Claude Mythos Preview, we investigated its internal mechanisms with interpretability techniques. We found it exhibited notably sophisticated (and often unspoken) strategic thinking and situational awareness, at times in service of unwanted actions. (1/14)
Image from Jack Lindsey (@Jack_W_Lindsey)
150
748
231
6,744
987,899
Runjin Chen retweeted
Anthropic (@AnthropicAI) profile picture
Anthropic
@AnthropicAIApr 7
Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software. It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans. anthropic.com/glasswing
Project Glasswing: Securing critical software for the AI era

Project Glasswing: Securing critical software for the AI era

A new initiative to secure the world’s most critical software and give defenders a durable advantage in the coming AI-driven era of cybersecurity.

anthropic.com
1,930
6,507
4,420
43,360
32,025,341
very happy to release this parameter generation work. from P-diff (2024), RPG (2025), DnD (2025) to HY-WU, parameter generation becomes more and more practical. thx @TencentHunyuan and happy to work with @oahzxl @mmbronstein @ZiqiaoWang63428 @cindy_x_wu @YangYou1991 @VITAGroupUT
Tencent Hy (@TencentHunyuan) profile pictureTencent Hy
@TencentHunyuanMar 5
One static model does not fit all😭 We just dropped our latest work: Functional Neural Memory. Instead of static models, we generate custom "parameters" for every single input. ✅Prompt your model anytime ✅Instant personalization ✅Better instruction following ✅Flexible & dynamic memory (w/o memory bank✌️) (🧵1/6)
4
8
35
5,362
🧠 Conclusion: Data curation is cognitive hygiene for AI. 🩺 Regular data “health checks” are essential for keeping models reliable, safe, and aligned. 💡 The striking parallels between AI and human cognitive decline may even offer new insights into human brain health. 👩‍🔬 Work by: Shuo Xing*, Junyuan Hong*, Yifan Wang, Runjin Chen, Zhenyu Zhang, Ananth Grama, Zhengzhong Tu, and Zhangyang Wang 🌐 Website (code, data, models): llm-brain-rot.github.io 📄 ArXiv: arxiv.org/abs/2510.13928
1
5
1
10
1,061
Runjin Chen retweeted
Ethan Perez (@EthanJPerez) profile picture
Ethan Perez
@EthanJPerez4 Sep 2025
We’re hiring someone to run the Anthropic Fellows Program! Our research collaborations have led to some of our best safety research and hires. We’re looking for an exceptional ops generalist, TPM, or research/eng manager to help us significantly scale and improve our collabs 🧵
10
40
12
258
71,373
In which the gang (@RunjinChen, @andyarditi, @Jack_W_Lindsey ): - identifies vectors for bad personas (evil, sycophancy, hallucinations, etc) - shows that if you inject the bad vectors in training, the model learns to not do the bad thing!! aka vaccines but for LLMs
Anthropic (@AnthropicAI) profile pictureAnthropic
@AnthropicAI1 Aug 2025
New Anthropic research: Persona vectors. Language models sometimes go haywire and slip into weird and unsettling personas. Why? In a new paper, we find “persona vectors"—neural activity patterns controlling traits like evil, sycophancy, or hallucination.
Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

ALT Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

4
9
4
92
10,555
Runjin Chen retweeted
Anthropic (@AnthropicAI) profile picture
Anthropic
@AnthropicAI1 Aug 2025
New Anthropic research: Persona vectors. Language models sometimes go haywire and slip into weird and unsettling personas. Why? In a new paper, we find “persona vectors"—neural activity patterns controlling traits like evil, sycophancy, or hallucination.
Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

ALT Our automated pipeline takes as input a personality trait (e.g. “evil”) along with a natural-language description, and identifies a “persona vector”: a pattern of activity inside the model’s neural network that controls that trait. Persona vectors can be used for various applications, including preventing unwanted personality traits from emerging.

219
869
314
5,705
1,430,966
Customizing Your LLMs in seconds using prompts🥳! Excited to share our latest work with @HPCAILab, @VITAGroupUT, @k_schuerholt, @YangYou1991, @mmbronstein, @damianborth : Drag-and-Drop LLMs(DnD). 2 features: tuning-free, comparable or even better than full-shot tuning.(🧵1/8)
5
74
11
112
18,681
Thrilled to share our latest project, "LLaGA: Large Language and Graph Assistant" 🚀 Dive into our findings here: arxiv.org/pdf/2402.08170.pdf. Plus, access our code on GitHub: github.com/VITA-Group/LLaGA
Image from Runjin Chen (@RunjinChen)
1
1
2
11
2,478
Key Feature: A versatile linear projector seamlessly bridges graph structures with the token space understood by Large Language Models (LLMs).
1
1
490
Our LLaGA excels in versatility, generalizability and interpretability, allowing it to perform consistently well across different datasets and tasks, extend its ability to unseen datasets or tasks, and provide explanations for graphs
1
375
Load more