Writer’s Post

This paper, titled “Chain of Thoughtlessness: An Analysis of CoT in Planning” dated 8 May 2024, presents a critical analysis of the effectiveness of the "Chain of Thought" (CoT) technique in enabling large language models (LLMs) to generalize reasoning abilities to out-of-distribution problems. The authors conduct a case study on using problems from the Blocksworld, a classical AI planning domain. They evaluated LLMs like GPT-4 and Claude-3-Opus across various CoT prompting strategies and problem complexities.   Context:   CoT prompting is a method of prompting LLMs to “reason” using intermediate steps before arriving at a final answer. The premise is that complex multi-steps problems could be tackled, when mimicking a train of thought, using carefully crafted prompts. Key Findings and Insights: ·     CoT prompts only provide meaningful performance improvements when they are exceedingly specific to the problem class being solved. ·     More general CoT prompts, which should theoretically enable LLMs to learn and apply algorithmic procedures, fail to provide robust reasoning capabilities across different problem instances. ·     As the complexity of problems increases, the benefits of CoT prompts quickly deteriorate, indicating a lack of true generalization. ·     Contrary to claims in the literature, the performance improvements from CoT prompts do not stem from LLMs learning general algorithmic procedures via examples of reasoning. ·     Highly specific CoT prompts can boost performance, but at the cost of significant human effort in crafting examples for each problem subclass. You can find the paper here:

A recent paper systematically tested prompt engineering strategies and defined 26 guiding principles for constructing prompts that lead to better performance from the LLMs. Performance was defined by two metrics: (1) boosting (human assessment of quality improvements) and (2) correctness, defined as outputs that are accurate, relevant, and error-free. It’s exciting to see systematic testing in the prompt engineering space, and these principles could act as an evaluation and testing framework when working with LLMs. If a prompt isn’t performing well, a first step could be to check it against these principles and implement any that are missing. Next, a similar experimental framework could be applied where the original prompt is compared with the new prompt using a validation dataset and metrics of interest. This paper is an encouraging signal of the growing interest in bringing systematic processes to working with LLMs. It is especially encouraging to us, as Openlayer was founded to bring clarity and ease to the AI evaluation process. As our focus has shifted to LLMs this past year, our urgency and dedication to our mission has only grown and we’re optimistic about the work that can be done to bring clarity to LLM evaluations. Read the paper here 👇 https://lnkd.in/dpYSspdB

📃Scientific paper: Universal Prompt Optimizer for Safe Text-to-Image Generation Abstract: Text-to-Image (T2I) models have shown great performance in generating images based on textual prompts. However, these models are vulnerable to unsafe input to generate unsafe content like sexual, harassment and illegal-activity images. Existing studies based on image checker, model fine-tuning and embedding blocking are impractical in real-world applications. Hence, we propose the first universal prompt optimizer for safe T2I (POSI) generation in black-box scenario. We first construct a dataset consisting of toxic-clean prompt pairs by GPT-3.5 Turbo. To guide the optimizer to have the ability of converting toxic prompt to clean prompt while preserving semantic information, we design a novel reward function measuring toxicity and text alignment of generated images and train the optimizer through Proximal Policy Optimization. Experiments show that our approach can effectively reduce the likelihood of various T2I models in generating inappropriate images, with no significant impact on text alignment. It is also flexible to be combined with methods to achieve better performance. Our code is available at https://lnkd.in/ejwPvHJ8. Continued on ES/IODE ➡️ https://etcse.fr/dRm4 ------- If you find this interesting, feel free to follow, comment and share. We need your help to enhance our visibility, so that our platform continues to serve you.

🚨Paper  Alert🚨 ➡️Paper Title:MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training 🌟Few pointers from the paper 🎯In this paper authors have discussed building performant Multimodal Large Language Models (MLLMs). 🎯 In particular, they studied the importance of various architecture components and data choices. Through careful and comprehensive ablations of the image encoder, the vision language connector, and various pre-training data choices, they identified several crucial design lessons. 🎯For instance in this paper authors have demonstrated that for large-scale multimodal pre-training using a careful mix of image-caption, interleaved image-text, and text-only data is crucial for achieving state-of-the-art (SOTA) few-shot results across multiple benchmarks, compared to other published pre-training results. 🎯 Further, the authors also showed that the image encoder together with image resolution and the image token count has substantial impact, while the vision-language connector design is of comparatively negligible importance. 🎯By scaling up the presented recipe, They built MM1, a family of multimodal models up to 30B parameters, consisting of both dense models and mixture-of-experts (MoE) variants, that are SOTA in pre-training metrics and achieve competitive performance after supervised fine-tuning on a range of established multimodal benchmarks. 🏢Organization: Apple 1️⃣Read the Full Paper here:https://lnkd.in/gMBNSzwX Find this Valuable 💎 ? ♻️REPOST and teach your network something new Follow me, Naveen Manwani, for the latest updates on Tech and AI-related news, insightful research papers, and exciting announcements. #multimodal #ai

🔍 Exciting Development in AI Reasoning: Meet RATIONALYST! 🔍 I came across a fascinating new model called RATIONALYST, designed to enhance reasoning capabilities in large language models (LLMs). 🤔 Often, the reasoning steps produced by LLMs can feel incomplete, mimicking the logical leaps we commonly make in everyday communication. To address this challenge, RATIONALYST introduces process-supervision for reasoning, pre-trained on a massive collection of rationale annotations. Here are some key highlights: - 79,000 rationales: Extracted from a vast web-scale dataset (The Pile) with minimal human intervention. - Versatile reasoning: RATIONALYST generalizes across diverse tasks, including - mathematical, commonsense, scientific, and logical reasoning. - Performance boost: Fine-tuned from LLaMa-3-8B, it improves accuracy by an average of 3.9% on seven representative reasoning benchmarks. - Strong competitor: RATIONALYST demonstrates superior performance compared to larger models like GPT-4! If you're curious to learn more about this innovative approach to improving reasoning in AI, check out the full paper here: https://lnkd.in/gGQSxuPp. #ai #machinelearning #naturallanguageprocessing #reasoning #huggingface #rationalyst

New Research paper for LLMs ♨ 𝗧𝗵𝗲 𝗘𝗿𝗮 𝗼𝗳 𝟭-𝗯𝗶𝘁 𝗟𝗟𝗠𝘀: 𝗔𝗹𝗹 𝗟𝗮𝗿𝗴𝗲 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝗿𝗲 𝗶𝗻 𝟭.𝟱𝟴 𝗕𝗶𝘁𝘀 https://lnkd.in/dBApMaMc We can't deny that Large language models (LLMs) have achieved state-of-the-art performance on a variety of tasks. However, LLMs can be computationally expensive and memory intensive, which limits their deployment on resource constrained devices. For example, a model like Megatron-Turing NLG which has 530B parameters. One way to reduce the computational cost and memory footprint of LLMs is to use lower-precision weights. For example, using 8-bit/4-bit weights can achieve a close performance to full-precision models with a significant reduction in memory usage. This research introduces "𝐁𝐢𝐭𝐍𝐞𝐭 𝐛𝟏.𝟓𝟖", a revolutionary LLM using 1-bit weights instead of the usual 16-bit!. Which can represent three values: -1, 0, or 1. This significantly reduces the memory footprint of the model compared to full-precision models. Their results shows that using the "𝐁𝐢𝐭𝐍𝐞𝐭 𝐛𝟏.𝟓𝟖" can achieve similar performance to the full-precision model on all tasks. They added some results for LLaMA 70B model using BitNet b1.58. 1. Throughput ~8.9x increase 2. Batch size ~11.0x increase 3. Energy consumption ~41.2x decrease 4. Decode latency ~4.1x decrease 5. Memory ~7.16x decrease Check out more in the research paper link! #ai #llm #researchpaper #machinelearning #largelanguagemodels

🚀 AI that plans and executes complex tasks like a strategic partner is no longer a far-off dream—it's the reality brought closer by a new AI breakthrough from Google's research team, presenting Language Agent Tree Search (LATS). 🌳✨ 🧠 LATS is a game-changer for businesses, as it equips language models with the unprecedented ability to reason, act, and plan. It's not just about answering queries anymore; it's about AI that can tackle decision-making tasks with human-like deliberation. Key Insights: 🔑 LATS integrates Monte Carlo Tree Search with LMs, enabling them to explore a variety of outcomes and make informed decisions. 🔑 It's a big leap forward, as abilities emerge in models with 100B+ parameters (current frontier models are already at this level!). 🔑 The framework uses an environment for feedback, allowing AI to adapt and solve problems more effectively. 🔑 In practice, LATS has shown impressive results, like a 92.7% accuracy in programming tasks with GPT-4. Analysis: 🔍 LATS could revolutionize industries by providing AI agents that can autonomously navigate complex environments and learn from their experiences. 🔍 For example, in e-commerce, LATS could autonomously navigate websites, compare products, and make purchases based on user-defined criteria. 🔍 While promising, it's crucial to consider the computational demands and the current necessity for environments that support state reversion. 🔗 Explore our in-depth analysis here: https://lnkd.in/eabPkgXg 🔎 How might LATS influence the evolution of autonomous systems in sectors beyond tech, such as healthcare or logistics? Would you trust an AI with strategic planning in your business? Let us know! #ArtificialIntelligence #MachineLearning #BusinessStrategy #Innovation

🌟 Exciting News! 🌟 Thrilled to share that I have seven papers accepted at NeurIPS 2024! 🎉 This year, our work spans several topics at the intersection of deep learning, optimization, and theoretical analysis. A special thanks to all my collaborators who contributed to these efforts. Here are the paper titles: 1️⃣ On the Comparison between Multi-modal and Single-modal Contrastive Learning 2️⃣ Provable and Efficient Dataset Distillation for Kernel Ridge Regression 3️⃣ Federated Learning from Vision-Language Foundation Models: Theoretical Analysis and Method 4️⃣ Unveil Benign Overfitting for Transformer in Vision: Training Dynamics, Convergence, and Generalization 5️⃣ Provably Transformers Harness Multi-Concept Word Semantics for Efficient In-Context Learning 6️⃣ On Mesa-Optimization in Autoregressively Trained Transformers: Emergence and Capability 7️⃣ SLTrain: a Sparse Plus Low Rank Approach for Parameter and Memory Efficient Pretraining Check out two of the arXiv versions here: https://lnkd.in/gjyrDnQt https://lnkd.in/gaBJ_4TD More of our work will be released soon, so stay tuned! 📚 #NeurIPS2024 #DeepLearning #AI #MachineLearning #Research

5 techniques to fine-tune LLMs, explained visually! Fine-tuning large language models traditionally involved adjusting billions of parameters, demanding significant computational power and resources. However, the development of some innovative methods have transformed this process. Here’s a snapshot of five cutting-edge techniques for finetuning LLMs, each explained visually for easy understanding. LoRA: - Introduce two low-rank matrices, A and B, to work alongside the weight matrix W. -Adjust these matrices instead of the behemoth W, making updates manageable. LoRA-FA (Frozen-A): - Takes LoRA a step further by freezing matrix A. - Only matrix B is tweaked, reducing the activation memory needed. VeRA: - All about efficiency: matrices A and B are fixed and shared across all layers. - Focuses on tiny, trainable scaling vectors in each layer, making it super memory-friendly. Delta-LoRA: - A twist on LoRA: adds the difference (delta) between products of matrices A and B across training steps to the main weight matrix W. - Offers a dynamic yet controlled approach to parameter updates. LoRA+: - An optimized variant of LoRA where matrix B gets a higher learning rate. This tweak leads to faster and more effective learning. Credits to Avi Chawla for great visualisation! 👏

𝐓𝐡𝐞 𝐏𝐫𝐨𝐦𝐩𝐭 𝐑𝐞𝐩𝐨𝐫𝐭: 𝐀 𝐒𝐲𝐬𝐭𝐞𝐦𝐚𝐭𝐢𝐜 𝐒𝐮𝐫𝐯𝐞𝐲 𝐨𝐟 𝐏𝐫𝐨𝐦𝐩𝐭𝐢𝐧𝐠 𝐓𝐞𝐜𝐡𝐧𝐢𝐪𝐮𝐞𝐬 This 76 page overview aims to provide a comprehensive taxonomy of prompting techniques. Presents 33 vocabulary terms, a taxonomy of 58 text-only prompting techniques and 40 techniques for other modalities as well as: 📝 Evaluated various prompting techniques on the MMLU benchmark. 📊 Discusses methods for evaluating prompt outputs to ensure accuracy and reduce risks. 🌍 Explores prompting techniques beyond English, including multilingual and multimodal approaches. 🎨 Techniques extend to various media like images, audio, and video, reflecting the growing complexity of prompting. 𝘋𝘚𝘗𝘺 𝘰𝘶𝘵𝘱𝘦𝘳𝘧𝘰𝘳𝘮𝘴 𝘢 𝘩𝘶𝘮𝘢𝘯 20:1 𝘢𝘯𝘥 𝘢𝘤𝘩𝘪𝘦𝘷𝘦𝘴 𝘣𝘦𝘵𝘵𝘦𝘳 𝘍1 𝘴𝘤𝘰𝘳𝘦: "We documented the prompt engineering process in order to illustrate the way that an experienced prompt engineer goes about their work. The exercise proceeded through 47 recorded development steps, cumulatively about 20 hours of work" Human prompting often involves trial and error to find effective templates. DSPy offers a more systematic approach by treating prompting as a program. By defining modules for steps like retrieving context and generating answers, DSPy helps build the prompt for you. Automated prompt engineering shows strong potential with DSPy. You will hear a lot more about DSPy over the next few months. Abs: https://lnkd.in/gjG4iPmD DSPy: https://lnkd.in/g8vw2Bc9 Github: https://lnkd.in/g7remNgw #AI #GenAI #LLM #MLLM #Prompting #Promptengineering #DSPY #multimodal