#what_i_read_today

# #what_i_read_today - [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://arxiv.org/abs/2310.03684) - [Dataset Condensation with Distribution Matching](https://arxiv.org/abs/2110.04181) - [Convolution for Computer Science People](https://medium.com/mlearning-ai/convolution-for-computer-science-people-2da7482272be) - [Score-Based Diffusion Models | Fan Pu Zeng](https://fanpu.io/blog/2023/score-based-diffusion-models/) - [NVIDIA LLM Developer AI Day](https://www.nvidia.com/en-us/events/llm-developer-day/?ncid=em-targ-521337) - [PoisonGPT: How to poison LLM supply chainon Hugging Face](https://blog.mithrilsecurity.io/poisongpt-how-we-hid-a-lobotomized-llm-on-hugging-face-to-spread-fake-news/) - [Hamming, "You and Your Research" (June 6, 1995)](https://www.youtube.com/watch?v=a1zDuOPkMSw) - [sigstore](https://www.sigstore.dev/) - https://www.lesswrong.com/posts/57fTWCpsAyjeAimTp/interpretability-in-ml-a-broad-overview-2 - https://arxiv.org/pdf/1611.03530.pdf - [GitHub - andyzoujm/representation-engineering: Representation Engineering: A Top-Down Approach to AI Transparency](https://github.com/andyzoujm/representation-engineering) - [Representation Engineering: A Top-Down Approach to AI Transparency](https://www.ai-transparency.org/) - [GitHub - guardrails-ai/guardrails: Adding guardrails to large language models.](https://github.com/guardrails-ai/guardrails) - [Guardrails AI | Your Enterprise AI needs Guardrails](https://docs.guardrailsai.com/) - [MetNet-3: A state-of-the-art neural weather model available in Google products](https://blog.research.google/2023/11/metnet-3-state-of-art-neural-weather.html) - [chiphuyen's list / Cool LLM repos](https://github.com/stars/chiphuyen/lists/cool-llm-repos) - [Idempotent Generative Network](https://arxiv.org/abs/2311.01462) - [PromptIDE](https://x.ai/prompt-ide/) - https://platform.openai.com/docs/models/gpt-4-and-gpt-4-turbo - [Adversarial Attacks on LLMs](https://lilianweng.github.io/posts/2023-10-25-adv-attack-llm/) - [Evaluation & Hallucination Detection for Abstractive Summaries](https://eugeneyan.com/writing/abstractive/) - https://x.com/0xgaut/status/1729177927445860374?s=20 - [What is a Vector Database & How Does it Work? Use Cases + Examples | Pinecone](https://www.pinecone.io/learn/vector-database/) - [Introducing Pika 1.0, An Idea-to-Video Platform](https://youtube.com/watch?v=6b10jGNNbXQ&si=mk853QDMV-IYNyuJ) - [My North Star for the Future of AI](https://www.theatlantic.com/technology/archive/2023/11/ai-ethics-academia/675913/?trk=feed-detail_main-feed-card_feed-article-content) - [Gemini - Google DeepMind](https://deepmind.google/technologies/gemini/#hands-on) - [Consistency Models](https://arxiv.org/abs/2303.01469) - [Gemini - Google DeepMind](https://deepmind.google/technologies/gemini/#hands-on…) - [The Gemini Lie](https://www.youtube.com/watch?v=90CYYfl9ntM) - [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://arxiv.org/abs/2312.00752) - https://x.com/sundarpichai/status/1734952757722001626?s=12 - https://t.co/ozfVwuBpSZ. - https://www.sciencedirect.com/science/article/pii/S0893608023006470?via%3Dihub#sec0013 - [Perspectives on the State and Future of Deep Learning - 2023](https://arxiv.org/abs/2312.09323) - [WhiteRabbitNeo/WhiteRabbitNeo-13B-v1 � Hugging Face](https://huggingface.co/whiterabbitneo/WhiteRabbitNeo-13B) - [Archives - colah's blog](https://colah.github.io/archive.html) - [Developing Llama 2 | Angela Fan](https://www.youtube.com/watch?v=NvTSfdeAbnU) - [Double descent - Wikipedia](https://en.m.wikipedia.org/wiki/Double_descent) - https://arxiv.org/pdf/1606.04838.pdf - [Highly accurate protein structure prediction with AlphaFold - Nature](https://www.nature.com/articles/s41586-021-03819-2) - [Broadly applicable and accurate protein design by integrating structure prediction networks and diffusion generative models](https://www.biorxiv.org/content/10.1101/2022.12.09.519842v2) - https://www.science.org/doi/10.1126/science.add2187 - [Highly accurate protein structure prediction with AlphaFold - Nature](https://www.nature.com/articles/s41586-021-03819-2) - [Aging with GRACE: Lifelong Model Editing with Discrete Key-Value Adaptors](https://arxiv.org/abs/2211.11031) - [skfolio](https://skfolio.org/) - [Optimize PyTorch Performance for Speed and Memory Efficiency (2022) | by Jack Chih-Hsu Lin | in Towards Data Science - Freedium](https://freedium.cfd/https://towardsdatascience.com/optimize-pytorch-performance-for-speed-and-memory-efficiency-2022-84f453916ea6) - [AlphaGeometry: An Olympiad-level AI system for geometry](https://deepmind.google/discover/blog/alphageometry-an-olympiad-level-ai-system-for-geometry/) - [The Faiss library](https://arxiv.org/abs/2401.08281) - [Generative Agents: Interactive Simulacra of Human Behavior](https://arxiv.org/abs/2304.03442) - [UniVTG: Towards Unified Video-Language Temporal Grounding](https://arxiv.org/abs/2307.16715) - [HOW HARD IS TROJAN DETECTION IN DNNS? FOOLING DETECTORS WITH EVASIVE TROJANS](https://openreview.net/pdf?id=V-RDBWYf0go) - [Do Explanations Reflect Decisions? A Machine-centric Strategy to Quantify the Performance of Explainability Algorithms](https://arxiv.org/abs/1910.07387) - [4 Autonomous AI Agents you need to know](https://towardsdatascience.com/4-autonomous-ai-agents-you-need-to-know-d612a643fa92) - [Image Restoration with Mean-Reverting Stochastic Differential Equations](https://arxiv.org/abs/2301.11699) - [CS25 I Stanford Seminar 2022 - Transformer Circuits, Induction Heads, In-Context Learning](https://youtu.be/pC4zRb_5noQ) - [LLM Attacks](https://github.com/llm-attacks/llm-attacks) - [AIM and continuous value data could transform computing](https://www.microsoft.com/en-us/research/blog/unlocking-the-future-of-computing-the-analog-iterative-machines-lightning-fast-approach-to-optimization/) - [Adversarial Examples Are Not Bugs, They Are Features](https://www.youtube.com/watch?v=hMO6rbMAPew) - [On Adaptive Attacks to Adversarial Example Defenses](https://arxiv.org/abs/2002.08347) - [Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples](https://arxiv.org/abs/1802.00420) - [Llama2](https://github.com/karpathy/llama2.c) - [Attention is Turing Complete](https://jmlr.org/papers/volume22/20-302/20-302.pdf) - [The Best Defense is a Good Offense: Adversarial Augmentation against Adversarial Attacks](https://arxiv.org/abs/2305.14188) - [Perspectives on diffusion](https://sander.ai/2023/07/20/perspectives.html) - [Deep Dive into Kernel Fusion: Accelerating Inference in Llama V2 - Lefebvre Sarrut's AI blog](https://ai.lefebvre-sarrut.eu/2023/07/20/deep-dive-into-kernel-fusion-accelerating-inference-in-llama-v2/) - [Yam Peleg on Twitter](https://twitter.com/Yampeleg/status/1680370128871989248?t=47bNYHO3XeRof2nfZm-FaQ&s=08) - [Matthias Niessner on Twitter](https://twitter.com/MattNiessner/status/1677715521951637507?t=uO9mEKeoWRvM2hwSrPlXkA&s=31) - [Keras: Deep Learning for humans](https://keras.io/keras_core/announcement/) - [Whose responsibility is responsible AI?](https://www.unusual.vc/post/responsible-ai) - [LLM trojan](https://huggingface.co/yifever/sleeper-agent) - [Tight Auditing of Differentially Private Machine Learning](https://arxiv.org/abs/2302.07956) - [Adversarial training and robustness for multiple perturbations](https://proceedings.neurips.cc/paper/2019/file/5d4ae76f053f8f2516ad12961ef7fe97-Paper.pdf) - [No Free Lunch in "Privacy for Free: How does Dataset Condensation Help Privacy"](https://arxiv.org/abs/2209.14987) - ["Real Attackers Don't Compute Gradients": Bridging the Gap Between Adversarial ML Research and Practice](https://arxiv.org/abs/2212.14315) - [Poisoning Web-Scale Training Datasets is Practical](https://arxiv.org/abs/2302.10149) - [Randomness in ML Defenses Helps Persistent Attackers and Hinders Evaluators](https://arxiv.org/abs/2302.13464) - [Label-Only Membership Inference Attacks](https://proceedings.mlr.press/v139/choquette-choo21a/choquette-choo21a.pdf) - [New ways of breaking app-integrated LLMs](https://github.com/greshake/llm-security) - [Drew Linsley on Twitter](https://twitter.com/DrewLinsley/status/1590763728009428992?t=-wlgIjpWsJRF4UFgUSgMFg&s=31) - [Grammatical Error Correction: Tag, Not Rewrite](https://www.grammarly.com/blog/engineering/gec-tag-not-rewrite/) - [How we built it: Stripe Radar](https://stripe.com/blog/how-we-built-it-stripe-radar) - [Inside GitHub: Working with the LLMs behind GitHub Copilot | The GitHub Blog](https://github.blog/2023-05-17-inside-github-working-with-the-llms-behind-github-copilot/) - [Reconstructing indoor spaces with NeRF](https://ai.googleblog.com/2023/06/reconstructing-indoor-spaces-with-nerf.html?m=1) - [Xerox scanners/photocopiers randomly alter numbers in scanned documents](https://www.dkriesel.com/en/blog/2013/0802_xerox-workcentres_are_switching_written_numbers_when_scanning) - [AGI safety difficult](https://twitter.com/mezaoptimizer/status/1667300224656715776) - [Nicolas Papernot on Twitter](https://twitter.com/NicolasPapernot/status/1664280922265616385?t=diOKd2NIkvXbxfE5XZcT-A&s=08) - [Navigating the Challenges of LLMs: Guardrails AI to the Rescue](https://mlsecops.com/podcast/shreya-navigating-the-challenges-of-llms-guardrails-ai-to-the-rescue) - [Parameter-Free Optimizers for Pytorch](https://github.com/bremen79/parameterfree) - [Sponge Examples: Energy-Latency Attacks on Neural Networks](https://arxiv.org/abs/2006.03463) - [Washing The Unwashable : On The (Im)possibility of Fairwashing Detection](https://proceedings.neurips.cc/paper_files/paper/2022/hash/5b84864ff8474fd742c66f219b2eaac1-Abstract-Conference.html) - [The Security Hole at the Heart of ChatGPT and Bing](https://www.wired.com/story/chatgpt-prompt-injection-attack-security/) - [Making LLMs even more accessible with bitsandbytes, 4-bit quantization and QLoRA](https://huggingface.co/blog/4bit-transformers-bitsandbytes) - [prompt injection in large language models](https://twitter.com/rharang/status/1659644874612953088) - [PyPI Repository Under Attack](https://thehackernews.com/2023/05/pypi-repository-under-attack-user-sign.html) - [Global and surrogate methods, interpretable models](https://www.borealisai.com/research-blogs/explainability-ii-global-explanations-proxy-models-and-interpretable-models/) - [Local post hoc methods](https://www.borealisai.com/research-blogs/explainability-i-local-post-hoc-explanations/) - [Writing Python like it’s Rust](https://kobzol.github.io/rust/python/2023/05/20/writing-python-like-its-rust.html) - [SAP/ml-model-watermarking](https://github.com/SAP/ml-model-watermarking) - [Where is the Information in a Deep Neural Network?](https://arxiv.org/abs/1905.12213) - [Confident Learning: Estimating Uncertainty in Dataset Labels](https://arxiv.org/abs/1911.00068) - [Compromised PyTorch Dependency Chain ](https://atlas.mitre.org/studies/AML.CS0015/) - [Machine Language Modelling from System Loggin](https://github.com/dtrizna/nebula) - [A tutorial on Differential Evolution with Python](https://pablormier.github.io/2017/09/05/a-tutorial-on-differential-evolution-with-python/) - [Faster Deep Learning Training with PyTorch – a 2021 Guide](https://efficientdl.com/faster-deep-learning-in-pytorch-a-guide/) - [Model Calibration](https://www.tidyverse.org/blog/2022/11/model-calibration/) - [Investigating the Nature of 3D Generalization in Deep Neural Networks](https://arxiv.org/abs/2304.09358) - [Transformers Agents](https://twitter.com/huggingface/status/1656334778407297027?t=y-vHjNxtRFVojB3zW7FTdA&s=31) - [ImageBind: One Embedding Space To Bind Them All](https://arxiv.org/abs/2305.05665) - [Unlimiformer: Long-Range Transformers with Unlimited Length Input](https://arxiv.org/abs/2305.01625) - [Product Launch 2023 Keynote](https://www.youtube.com/watch?v=-3Kf2ZZU-dg) - [HuggingChat](https://twitter.com/ClementDelangue/status/1650908484936908808?t=xF4KewK3S3TEnQi5yeP5qA&s=19) - [Meta AI on Semi Supervised Learning](https://twitter.com/ylecun/status/1650798206283051009?s=20) - [NeRFs on Google Search](https://twitter.com/rmbrualla/status/1649270583782412293) - [Interpretability of Transformers with up to two layers of attention](https://transformer-circuits.pub/2021/framework/index.html) - [Using Softmax Linear Units(SoLU) to investigate interpretability of transformers](https://transformer-circuits.pub/2022/solu/index.html) - [Beyond automatic differentiation](https://ai.googleblog.com/2023/04/beyond-automatic-differentiation.html?m=1) - [Cultivating Your Research Taste](https://medium.com/great-research/cultivating-your-research-taste-ce77bbee7f2f) - [Choose Your Weapon: Survival Strategies for Depressed AI Academics](https://arxiv.org/abs/2304.06035) - [Approximating Wasserstein distances with PyTorch](https://dfdazac.github.io/sinkhorn.html) - [Adversarial Data Augmentation with Chained Transformations (AdvChain)](https://twitter.com/cherise_go/status/1644031218583760918?t=cj78D6NtnjDyx-pENn9nrA&s=31) - [ModelDiff: A Framework for Comparing Learning Algorithms](https://twitter.com/aleks_madry/status/1595512357140455424?t=xgeAsvcttrjse-h3_Gjg5A&s=31) - [Stochastic Weight Averaging — a New Way to Get State of the Art Results in Deep Learning](https://pechyonkin.me/stochastic-weight-averaging/) - [30B model now needs only 5.8GB of RAM? How?](https://github.com/ggerganov/llama.cpp/discussions/638) - [Ilya Sutskever (OpenAI Chief Scientist) - Building AGI, Alignment, Spies, Microsoft, & Enlightenment](https://www.youtube.com/watch?v=Yf1o0TQzry8) - [Watermarking for Out-of-distribution Detection](https://arxiv.org/abs/2210.15198) - [Continual Few-Shot Learning Using HyperTransformers](https://arxiv.org/abs/2301.04584) - [NeurIPS 2022 Workshop MLSW Submissions](https://openreview.net/submissions?venue=NeurIPS.cc/2022/Workshop/MLSW) - [Creating Confidence Intervals for Machine Learning Classifiers](https://sebastianraschka.com/blog/2022/confidence-intervals-for-ml.html) - [26ms Inference Time for ResNet-50: Towards Real-Time Execution of all DNNs on Smartphone](https://arxiv.org/abs/1905.00571) - [tinygrad: A simple and powerful neural network framework](https://tinygrad.org/) - [ GPT in 60 Lines of NumPy | Jay Mody](https://jaykmody.com/blog/gpt-from-scratch/) - [System 2 Is What We Need](https://system2.ai/p/system-2-is-what-we-need) - [Quick tour - BlindLlama](https://blindllama.mithrilsecurity.io/en/latest/docs/getting-started/quick-tour/?pk_campaig[…]-11-2023dhlkLaunch_Blindlama&pk_source=Lk&pk_medium=SN) - [Validating LLM Outputs](https://txt.cohere.com/validating-llm-outputs/) - [The Rise and Potential of Large Language Model Based Agents: A Survey](https://arxiv.org/abs/2309.07864) - https://twitter.com/norabelrose/status/1701596533756485901?t=O9PcqotmxbWfYUXeMH0IDw&s=08 - [Slack](https://dsgiitr.slack.com/archives/C05SLLQGFTP/p1694948384487309) - [GitHub - laiyer-ai/llm-guard: The Security Toolkit for LLM Interactions](https://github.com/laiyer-ai/llm-guard) - [Laiyer: Unleash LLM�s potential with confidence](https://laiyer.ai) - [Introduction to AI Accountability & Transparency Series](https://aipolicy.substack.com/p/ai-accountability-transparency-intro) - [FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance](https://arxiv.org/abs/2305.05176) - https://twitter.com/cHHillee/status/1704140759224418676?t=4qBbqpf6UFg3iQikYqADEg&s=19 - [Visualizing PyTorch memory usage over time](https://t.co/3s61R4UYBU) - [A tale of two problem solvers (Average cube shadows)](https://youtu.be/ltLUadnCyi0?si=TlvQxud8mNzhbfLP) - [From Newton�s method to Newton�s fractal (which Newton knew nothing about)](https://youtu.be/-RdOwhmqP5s?si=uZ6ET0DGTm2d5a2a) - [Full Event | #MicrosoftEvent September 21, 2023](https://www.youtube.com/watch?v=XYUEQ0SyOyE) - [The Adventure of the Errant Hardware](https://www.adept.ai/blog/sherlock-sdc) - [Writing Python like it�s Rust](https://kobzol.github.io/rust/python/2023/05/20/writing-python-like-its-rust.html) - https://www.marktechpost.com/2023/09/21/ibm-researchers-propose-a-new-adversarial-attack-framework-capable-of-generating-adversarial-inputs-for-ai-systems-regardless-of-the-modality-or-task/?amp - [A Practical Deep Learning-Based Acoustic Side Channel Attack on Keyboards](https://arxiv.org/abs/2308.01074) - https://twitter.com/abhishekunique7/status/1699263297252377037 - https://twitter.com/abhi9u/status/1707375199233151197?t=VPIUmADPFNMkxU4IfUNxMg&s=19 - [GitHub - guardrails-ai/guardrails: Adding guardrails to large language models.](https://github.com/ShreyaR/guardrails) - [Guardrails AI | Your Enterprise AI needs Guardrails](https://docs.guardrailsai.com/) - [Dataset Condensation with Distribution Matching](https://arxiv.org/abs/2110.04181) - [AutoPrompt: Eliciting Knowledge from Language Models with Automatically Generated Prompts](https://arxiv.org/abs/2010.15980) - [AutoPrompt](https://ucinlp.github.io/autoprompt/) - [Representation Engineering: A Top-Down Approach to AI Transparency](https://www.ai-transparency.org/) - [Work & Projects Summary | CAIS](https://www.safe.ai/work-summary) - [Scaling up learning across many different robot types](https://www.deepmind.com/blog/scaling-up-learning-across-many-different-robot-types) - [Rewind Pendant](https://www.rewind.ai/pendant) - [NVIDIA Technical Blog | News and tutorials for developers, data scientists, and IT admins](https://developer.nvidia.com/blog) - https://openreview.net/pdf?id=bYV3bK_Azi - [Google Colaboratory](https://colab.research.google.com/drive/1F6_1_cWXE5M7WocUcpQWp3v8z4b1jL20) - [Neel Nanda](https://www.neelnanda.io/) - [Generating Synthetic Dataset for RAG � Nextra](https://www.promptingguide.ai/applications/synthetic_rag) - [The AI research job market shit show (and my experience)](https://www.interconnects.ai/p/ai-research-job-market?utm_content=buffer7f9a8&utm_medium=social&utm_source=linkedin.com&utm_campaign=buffer) - https://arxiv.org/pdf/1906.08988.pdf - [keerthanapg](https://keerthanapg.com/) - [AI's Underbelly: The Zero-Day Goldmineby: Dan McInerney](https://www.youtube.com/watch?v=e3ybnXjtpIc) - [huntr - The world�s first bug bounty platform for AI/ML](https://huntr.mlsecops.com/bounties) - [GitHub - jxmorris12/vec2text: utilities for decoding deep representations (like sentence embeddings) back to text](https://github.com/jxmorris12/vec2text/) - [Compiling NumPy code into C++ or CUDA via torch.compile](https://pytorch.org/blog/compiling-numpy-code/?utm_content=268301731&utm_medium=social&utm_source=linkedin&hss_channel=lcp-78618366) - [SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks](https://arxiv.org/abs/2310.03684) - [Dataset Condensation with Distribution Matching](https://arxiv.org/abs/2110.04181) - [Convolution for Computer Science People](https://medium.com/mlearning-ai/convolution-for-computer-science-people-2da7482272be) - [Score-Based Diffusion Models | Fan Pu Zeng](https://fanpu.io/blog/2023/score-based-diffusion-models/) - [PoisonGPT: How to poison LLM supply chainon Hugging Face](https://blog.mithrilsecurity.io/poisongpt-how-we-hid-a-lobotomized-llm-on-hugging-face-to-spread-fake-news/) - [Hamming, "You and Your Research" (June 6, 1995)](https://www.youtube.com/watch?v=a1zDuOPkMSw) - [sigstore](https://www.sigstore.dev/) - https://www.lesswrong.com/posts/57fTWCpsAyjeAimTp/interpretability-in-ml-a-broad-overview-2 - https://arxiv.org/pdf/1611.03530.pdf - [GitHub - andyzoujm/representation-engineering: Representation Engineering: A Top-Down Approach to AI Transparency](https://github.com/andyzoujm/representation-engineering) - [Representation Engineering: A Top-Down Approach to AI Transparency](https://www.ai-transparency.org/)

Syntax	Example	Reference
# Header	Header	基本排版
- Unordered List	Unordered List
1. Ordered List	Ordered List
- [ ] Todo List	Todo List
> Blockquote	Blockquote
Bold font	Bold font
Italics font	Italics font
~~Strikethrough~~	~~Strikethrough~~
19^th^	19^th
H~2~O	H₂O
++Inserted text++	Inserted text
==Marked text==	Marked text
[link text](https:// "title")	Link
![image alt](https:// "title")	Image
`Code`	`Code`	在筆記中貼入程式碼
```javascript var i = 0; ```	`var i = 0;`
:smile:		Emoji list
{%youtube youtube_id %}	Externals
$L^aT_eX$	L^aT_eX
:::info This is a alert area. :::	This is a alert area.