Ignacio Martinez

Stole a few points from Prof Christopher Manning, who talked about LLMs and language modeling in general in the latest TWIML AI Podcast: 📌 Humans acquire language skills in a way very different from LLMs. We need millions of words, compared to billions or even trillions of tokens for LLMs. LLM researchers may want to investigate and learn from how humans acquire language skills. 📌 LLMs cannot reason. However, there are other deep learning models, such as AlphaGo, that can. LLM researchers may want to look into how to integrate that type of reasoning/searching/planning capability into LLMs. 📌 LLMs' world models should enable search and discovery. Although Prof. Manning didn’t call this out explicitly, my understanding is more similar to a knowledge graph type of structure. 📌 Next-gen LLM idea: a soft form of locality and hierarchy. Transformers attend every token to every other token, which is very inefficient, while human language can be modeled by n-grams most of the time. #artificialintelligence #machinelearning #deeplearning https://lnkd.in/euzwMQ6p

84

18 Comments

Open-source AI for evaluating language models Researchers from KAIST AI, LG AI Research, CMU, MIT, Allen Institute for AI, and UIC have introduced Prometheus 2, a groundbreaking open-source evaluator model that assesses language model outputs with accuracy rivaling expensive proprietary solutions like GPT-4. A few key highlights: Prometheus 2 combines models trained on direct assessment and pairwise ranking, enabling it to excel at both common evaluation formats. It leverages the new Preference Collection dataset with 1,000 evaluation criteria to boost performance. Prometheus 2 achieved the highest correlation with human judgments and proprietary models across 8 benchmarks, in some cases reaching 0.90 Pearson correlation and 85%+ accuracy. This helps narrow the gap between open-source and closed-source AI evaluation without sacrificing transparency and control As an AI and Data Strategist, I like that this will make high-quality AI evaluation more accessible and scalable for researchers and practitioners. Some potential strategic impacts: Accelerating AI/ML system development by enabling rapid, reliable evaluation of generated language outputs at lower cost Establishing clearer standards and benchmarks for assessing language model performance as open alternatives mature #ai #machinelearning #nlp #opensource #research #aistrategy Paper - https://lnkd.in/gsnGDwJd

1

Very insightful work by Anthropic’s interpretability team. And an amazing paper, with outstanding writing and figures. The idea is very simple: interpret LLMs by leveraging sparse autoencoders as surrogate models of the MLP of transformer blocks, which allow one to disambiguate the superposition of features captured by a single neuron. A simple idea, but a very careful and complex execution, as it is often the case in our line of work. The paper goes into many details and provide a large array of insights, although the gist of the implementation remains obfuscated due to the closed source nature of Claude. Too bad, because this is the kind of work that we need to better understand and eventually trust LLMs. This is demonstrated by the authors in the section ‘Influence on Behavior’, where they show that clamping some features to either high or low value during inference is “remarkably effective at modifying model outputs in specific, interpretable ways”. Hopefully this kind of work is going to be replicated and generalized to open-weights models, such that we have new ways to steer their behavior. https://lnkd.in/eVym7f_f #interpretability #xai #explainableai #steerableai #anthropic #claude #anthropic

3

one of the challenges with RAG is evaluating it. check out this code from Amazon Science if you are building a RAG based system Romina Sharifpour, Ph.D. Baichuan Sun

23

5 Comments

After reading Anthropic and OpenAI success stories with sparse autoencoders for interpretability of text, I was curious about its application to images... After a quick search, I found out that Hugo Fry wrote a blogpost in LessWrong: https://lnkd.in/dA33fWqE which also provides some playground! Of course part of the success is due to the ImageNet structure itself (as many papers out there) but still is an interesting exercise. It enables lots of experiments, and I already can think about practical usage for improving image search in marketplaces. #interpretability #image #CLIP

7

3 Comments

One of the biggest challenges in language models today is making them more interpretable. We often treat AI models as black boxes: data goes in, a response comes out, and the reasoning behind that response remains unclear. I remember an interview with Google CEO, where he was asked to explain how Gemini works. He said he didn’t know. This response resonated with the scientific community, as deep learning often similar to the human brain, but the interviewer was shocked! How can a model, released to millions, be so poorly understood? Two weeks ago, Anthropic released an important paper on model interpretability. They used a technique called "dictionary learning," borrowed from classical ML, which isolates patterns of neuron activations that recur across many different contexts. This paper sheds some light on this important challenge, which, if solved, will create more trust in these models and thus ease the integration of AI into our everyday lives. Highly recommend reading: https://lnkd.in/gPzEePx8

12

1 Comment

New Berkeley on understanding large language model: foundations and safety. Lectures on YouTube and covers below topics. https://lnkd.in/g5ukTUV2

9

Hallucinations in LLMs are inevitable, no seriously! Josh Fritzsche and I have worked hard at Voze to find a way to tune, prompt and even plead 🙏 with LLMs to prevent them from the occasional 🍄 hallucination. It seemed like the better we have done the harder the edge cases have been to solve and eventually we realized that they tend to frequently happen when LLMs predict or make anticipations on "nothing". I remember at one point nearly going insane with frustration because when LLMs had a hallucination, they were EPIC! 😱 Then we pushed through it and benefited for the effort! How did we solve this common problem? We didn't (Not completely), we got better at identifying the issues as they happened. Once we identified in the data we were able to start looking for patterns. The same catalysts that cause it to happen once if left unchanged caused it over and over again. Our solution is astoundingly simple. Once you know that it's likely to happen..... let it! Accept it will 🍄, but also be preemptive in not focusing on it and use alternative data sources, generators or providers. Change the conditions that cause it but continue to monitor it as it happens. As Dan Caffee is fond of saying to me on a regular bases. "Dusty we need the system to be adaptive and provide mechanisms to learn from it's past." Well he is right! All we needed was data and the people that contribute to it's quality. Huge props 👏 to strong support from Kathy and Janelles audit teams. We were armed with a knowledge and a desire to learn from past mistakes. Armed with confidence that we could correct these problems we could focus on the data to prevent their impact going forward. When doing millions of LLM (Generative AI) calls regularly Josh and I have learned that hallucinations will happen. We also know exactly how we will handle and prevent it going forward! Failure is not doing anything and accepting defeat.

4

2 Comments

🚀 We are thrilled to announce the release of our new paper: "The Evolution of Multimodal Model Architectures." In this paper, we provide a bird's-eye view of all the state-of-the-art multimodal (large) models that have emerged since the inception of Transformers. It highlights the pros and cons of each model architecture, grouping them into four broad categories. These categories can serve as a blueprint to help you determine which architecture is best suited for your specific scenario. A huge thank you to all the co-authors and especially to Shakti Wadekar for putting this paper together during his PhD at Purdue University. Paper: 🔗 [https://lnkd.in/gZHQUS6c] Authors: Shakti Wadekar, Abhishek Chaurasia, Aman Chadha, and Eugenio Culurciello #AI #MachineLearning #MultimodalModels #Transformers #Research #PurdueUniversity

76

4 Comments

If you're interested in understanding the inner layers of LLM models, the #Anthropic blog is amazing. They tore down the Claude 3.0 Sonnet LLM models apart and peeked into its model view and perspectives. It's a long read paper, but an amazing read! By successfully extracting millions of features from the middle layer of their Claude 3.0 Sonnet model, they have uncovered a conceptual map of its internal representations, revealing how it encodes diverse concepts like cities, scientific fields, and even abstract notions supporting the security, various bias and power-seeking behavior etc... Good Read for a long weekend. https://lnkd.in/gNm8qA3W

7

2 Comments

This is guide for any AI Engineer looking to build autonomous AI Agents. #llms #chatgpt4 #llmops #ai #datascience #mlopd #aiagent https://lnkd.in/g4md6Vvn

4

Spatial Reasoning did not ‘emerge’ spontaneously in Large Language Models (LLMs) the way so many reasoning capabilities did. Humans have specialized, highly capable spatial reasoning capabilities that LLMs have not replicated. But every subsequent release of the major models — GPT, Claude, Gemini- promises better multimedia support, and all will accept and try to use uploaded graphics along with texts. #llms #chatgpt4 #gemini #llmops #aiml #ml https://lnkd.in/dFSWKqKs

2

📣 📣 📣 [CFP; MRR 2024 Workshop @ SIGIR 2024 Conference]We're excited to invite researchers and practitioners across the spectrum of multimodal modeling, representation learning, and retrieval  to participate in our workshop on "Multimodal Representation and Retrieval." This event is proudly organized in association with SIGIR 2024. Multimodal data is available in many applications like e-commerce production listings, social media posts and short videos. However, existing algorithms dealing with those types of data still focus on uni-modal representation learning by vision-language alignment and cross-modal retrieval. In this workshop, we target to bring a new retrieval problem where both queries and documents are multimodal. With the popularity of vision language modeling,  large language models (LLMs), retrieval augmented generation (RAG), and multimodal LLM, we see a lot of new opportunities for multimodal representation and retrieval tasks. This event will be a comprehensive half-day workshop focusing on the subject of multimodal representation and retrieval. The agenda includes keynote speeches, oral presentations, and an interactive panel discussion. Our objective with this workshop is to capture the interest of researchers in this emerging challenge. By highlighting the novelty and significance of the problem, we aim to attract researchers who are eager to explore and contribute to this field. Join us as we welcome esteemed speakers including Tat-Seng Chua (National University of Singapore), and Dinesh Manocha (University of Maryland), who will share their invaluable insights and expertise. We invite original research & industrial application papers that present research on multimodal data representation and retrieval. 🔗 Workshop website: https://lnkd.in/gcpZtrdc 📅 Submission Deadline: May 5, 2024 (11:59 pm, AOE) 📅 Workshop Date: July 18th, 2024 🌆 Workshop Location: Washington D.C., USA Join us to shape the future of multimodal representation, retrieval, and related applications! ☀ My co-organizers Arnab Dhua (Amazon), Douglas Gray (Amazon), I. Zeki Yalniz (Meta), Tan Y. (Nvidia), Mohamed Elhoseiny (KAUST), Bryan A. Plummer (Boston University) and I look forward to hosting you.

73

3 Comments

Prompt engineering is a rapidly-evolving topic in AI, but recent research can be grouped into four categories... (1) Reasoning: Simple prompting techniques are effective for many problems, but more sophisticated strategies are required to solve multi-step reasoning problems. - [1] uses zero-shot CoT prompting to automatically generate problem-solving rationales to use for standard CoT prompting. - [2] selects CoT exemplars based on their complexity (exemplars that have the maximum number of reasoning steps are selected first). - [3] improves CoT prompting by asking the LLM to progressively refine the generated rationale. - [4] decomposes complex tasks into several sub-tasks that can be solved via independent prompts and later aggregated into a final answer. (2) Tool Usage: LLMs are powerful, but they have notable limitations. We can solve many of these limitations by teaching the LLM how to leverage external, specialized tools. - [5, 6] finetune a language model to teach it how to leverage a fixed, simple set of text-based APIs when answering questions. - [7] uses a central LLM-based controller to generate a program—written in natural language—that composes several tools to solve a complex reasoning task. - [8] uses a retrieval-based finetuning technique to teach an LLM to adaptively make calls to APIs based on their documentation when solving a problem. - [9] uses an LLM as a central controller for leveraging a variety of tools in the form of deep learning model APIs. - [10, 11] integrates code-capable LLMs with a sandboxed Python environment to execute programs when solving problems. (3) Context Window: Given the emphasis of recent LLMs on long contexts for RAG / few-shot learning, the properties of context windows and in-context learning have been studied in depth. - [12] shows that including irrelevant context in the LLM’s prompt can drastically deteriorate performance. - [13] finds that LLMs pay the most attention to information at the beginning/end of the prompt, while information placed in the middle of a long context is forgotten. - [14] proposes a theoretically-grounded strategy for optimally selecting few-shot exemplars. (4) Better Writing: One of the most popular use-cases of LLMs is for improving human writing, and prompt engineering can be used to make more effective writing tools with LLMs. - [15] improves the writing abilities of an LLM by first generating an outline and then filling in each component of the outline one-by-one. - [16] uses a smaller LLM to generate a “directional stimulus” (i.e., a textual hint) that can be used as extra context to improve an LLM’s writing ability on a given task. - [17] improves the quality of LLM-generated summaries by iteratively prompting the LLM to increase the information density of the summary.

221

8 Comments

Looking how to select better representatives from your clusters? Take a look at Lexin Chen's PRIME repo: https://lnkd.in/enM_Btbx Now with updated functionality and tutorials! Great collaboration with Arup Mondal and Alberto Perez! #moleculardynamics #clustering #machinelearning

65

Large language models (LLMs) are known to generate plausible sounding factually incorrect responses, especially when faced with unfamiliar concepts, queries. Recent research suggests that strategic fine-tuning and in-context learning, retrieval augmentation can significantly reduce these hallucinations. #llms #chatgpt4 #llmops #aiml #gemini #phi #mlops https://lnkd.in/duRrC3yg

4