Paresh Nakhe

Can LLM preference tuning degrade Multi-Agent system performance? In a recent tweet, Andrej Karpathy questioned why all LLMs sound similar, prompting Sebastian Rascha to speculate about the uniformity of alignment datasets. Aligning user preferences is a last-mile step in training large language models. Regardless of the pre-training datasets used, could the preference datasets used during alignment compromise the effectiveness of LLMs? This issue could adversely affect multi-agent systems incorporating multiple LLMs in their design.

7

1 Kommentar

This is really awesome. I used notebookLM to summarise a recent paper that shows how o1 like model can be reproduced. After providing the pdf from Arxiv, I noticed that the generated podcast was almost 49 minutes long. Very different from the earlier 10-12 minute version. And another impressive feature was the interactive mode where you can interject the two hosts and ask a question which they very graciously respond to. I wanted to share the audio summary but YT didn’t allow me to upload the WAV file. YT expects a video mp4 file. So I found free videos from Pexels and stitched them together with ffmpeg. The videos are meditative acquarium shots of fish swimming around. And with the audio discussing reproducing o1 as an engaging podcast episode, the video keeps the distraction to the minimum. The paper itself is a great read but if you have a 30-40 minute commute, or want to watch some cool fish swimming around and learn about o1, check out

12

Today's reading is Memory Layers in Large Language Models 🧠 This research shows how memory layers can take LLM performance to the next level: 🔑 Trainable key-value lookup mechanism: Adds parameters without increasing computational cost, boosting factual accuracy and task performance. 📈 Big results, small cost: Outperforms models with much larger computational budgets and even beats mixture-of-experts models! ⚙️ Scales seamlessly with up to 128 billion memory parameters The authors from Meta highlights how memory layers are a game-changing addition to future AI architectures, offering smarter, faster, and more accurate solutions. You may want to check out the paper at: https://lnkd.in/gzuqUfYe #AI #LLM #MemoryLayers #Innovation #MachineLearning #FactualAI #FutureOfAI #Meta

19

"𝘕𝘰𝘣𝘰𝘥𝘺 𝘸𝘪𝘭𝘭 𝘯𝘦𝘦𝘥 𝘵𝘰 𝘤𝘰𝘥𝘦 𝘴𝘰𝘰𝘯!", right? Wrong. Oversimplified. Populistic. Hear me out: 𝗚𝗲𝗻𝗔𝗜 𝘄𝗶𝗹𝗹 𝗺𝗮𝗸𝗲 𝗴𝗼𝗼𝗱 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝘀 𝗯𝗲𝘁𝘁𝗲𝗿, 𝗮𝗻𝗱 𝗯𝗮𝗱 𝗰𝗼𝗱𝗲𝗿𝘀 𝘄𝗼𝗿𝘀𝗲. Want to become really good at using genAI? You need to understand a programmer's way of thinking: loops, algorithms, conditionals, DRY, KISS, data structures and then some... Why? 🤖 AI needs clear modular and often atomic instructions. 🛠️ Debugging and refinement are still essential. 💡 Innovation comes from combining domain expertise with this mindset. This isn't about becoming a coder, it is about becoming a better problem solver in the age of AI. Problem solving is what engineering is all about. And learning to code is one of the easiest ways to thinking like an engineer. WDYT? Let me know below!

19

2 Kommentare

[research update] use a Transformer's multi-head attention activations to fetch RAG style, to augment multi faceted problem breakdown for LLMs in ."Multi-Head RAG: Solving Multi-Aspect Problems with LLMs": 1. Enhances LLMs by integrating document retrieval (RAG) into the LLM context. 2. Challenge: Existing RAG solutions struggle with queries needing multiple documents with varied contents due to distant embeddings. 3. Multi-Head RAG (MRAG): Introduces a novel approach using Transformer's multi-head attention activations to fetch diverse documents. 4. Activation Utilization: Different attention heads capture varied data aspects, improving complex query retrieval. 5. Improved Relevance: MRAG shows up to 20% improvement in relevance over standard RAG baselines. 6. Evaluation Methodology: Provides metrics, synthetic datasets, and real-world use cases to demonstrate effectiveness. 7. Seamless Integration: MRAG can be integrated with existing RAG frameworks and benchmarking tools like RAGAS. 8. Multi-Aspect Problem Solving: Designed specifically to handle multi-aspect queries more effectively. 9. Data Stores Compatibility: Compatible with various data store classes. 10. Potential Applications: Benefits scenarios requiring diverse information retrieval, enhancing LLM responses. Experiment The paper introduces MRAG and evaluates it using synthetic datasets and real-world data. It compares MRAG's performance against standard RAG models. The experiment focuses on the model's ability to retrieve relevant documents across various query aspects, demonstrating up to a 20% improvement in relevance over baselines. Use Case MRAG is tested on real-world scenarios like customer support, where queries often require information from multiple documents with diverse contents. The model's capability to retrieve and integrate multi-aspect data enhances the quality of responses.

81

5 Kommentare

LLM as a solver Software involving Combinatorial optimization techniques is ubiquitous. How does LLM fare compare to a brute force method? We prompt the LLM with, 1. A distance matrix, the distance between the cities 2. Starting node/city. And expect the shortest path between the cities starting from the given start city and the distance covered. https://lnkd.in/enMkT92W 1. Create a python dataclass to describe the input and out of a TSP problem. 2. Write a prompt template to describe the input, output and few-shot examples. 3. A quick brute force TSP problem generator for n-shot examples. 4. Prompt the LLM for answer.

17

4 Kommentare

OpenCoder doesn't get enough love They open-sourced the entire pipeline to create QwenCoder-level code models. This includes: - Large datasets - High-quality models - Eval framework Tons of great lessons and observations in the paper 📝 Paper: arxiv.org/abs/2411.04905 #special #linkedin #Family #Pakistan #skills #islam #government #AI #ML #MachineLearning #tesla #microsoft #openai

5

Tomorrow is the last day to get substantially reduced early bird tickets for AMLD - Applied Machine Learning Days 2025. Here's my ❤️ pitch why it's the #AI conference you cannot miss: 🎯 The content: Almost 60 tracks and hands-on workshops. The tracks bring you face-to-face with the movers and shakers actively shaping AI's future, while the workshops get you coding and building. This year's quality bar was the highest we've seen in nearly a decade! 🌍 The audience: Our audience brings together decision-makers from across the AI landscape: engineers, scientists, business leaders, practitioners, policy makers, educators, entrepreneurs. This diversity creates unique networking opportunities - whether you're looking for collaborators, talent, investors, or clients, they'll be there. The quality of our audience is why so many participants return year after year. 🤝 The unique mix: AMLD is neither a pure scientific meeting nor trade show - it's a unique mix of both. Having started this at EPFL, one of the leading technical universities, we wanted to create something different: a place where breakthrough research meets practical implementation. Countless projects and collaborations have literally started at AMLD. This mix simply works. 🏔️ The setting: The SwissTech Convention Center is simply extraordinary. And when you step out, you're just minutes away from Lac Léman, with breathtaking views of *the* highest peak of the Alps. EPFL's campus has this unique energy, and being here in February means you can plan to hit the slopes, which should be amazing this time of the year!⛷️ ⚡ The timing: Everyone now realizes that AI is reshaping everything. At AMLD, you can go beyond the social media noise to meet and learn from the people who are actually building and deploying these systems. Whether you're deep in the technical challenges or leading strategic decisions: AMLD will be time well spent. Early bird tickets available until tomorrow at https://lnkd.in/exgqpKah AMLD - Applied Machine Learning Days EPFL AI Center EPFL Alumni EPFL School of Computer and Communication Sciences EPFL School of Life Sciences EPFL School of Engineering

85

🔮RAG + KG: Powerful Combination for Translation Multi-AI Agents Large language models (LLMs) have shown great promise in machine translation, but they still struggle with contextually dependent terms, such as new or domain-specific words. This leads to inconsistencies and errors that are difficult to address. Existing solutions often depend on manual identification of such terms, which is impractical given the complexity and evolving nature of language. While RetrievalAugmented Generation (RAG) could provide some assistance, its application to translation is limited by issues such as hallucinations from information overload. =================== In this paper, researchers propose CRAT, a novel multi-agent translation framework that: 》 leverages RAG and causality enhanced self-reflection to address these challenges. This framework consists of several specialized agents: ✱ the Unknown Terms Identification agent detects unknown terms within the context, ✱ the Knowledge Graph (KG) Constructor agent extracts relevant internal knowledge about these terms and retrieves bilingual information from external sources, ✱ the Causality enhanced Judge agent validates the accuracy of the information, ✱ Translator agent incorporates the refined information into the final output. =================== This automated process allows for more precise and consistent handling of key terms during translation. Their results show that CRAT significantly improves translation accuracy, particularly in handling context-sensitive terms and emerging vocabulary. --- If your company is looking for an AI consultant to accelerate business growth, or if you are looking to build your career in AI, send me a message and I can help

4

Relevant context is more important than long context, just as relevant data is more important than a large dataset. The effectiveness of relevant document retrieval is a key step. This boils down to the effectiveness of search and match (similarity metrics). It would be great, we understand the black box before using it. #rag #generativeAI

3

I'll never be able to read the word "delve" again without immediately thinking: Ah, this is AI writing. Anyway! A new paper has been published, along with fine-tuned models for generating really long outputs. The researchers weren't satisfied with the fact that current LLMs could technically produce much more output than they normally do. ChatGPT and similar models typically generate around 2000 words, despite being capable of producing far more tokens. Their hypothesis: The models don't encounter longer outputs during training. Initially, the researchers used prompt chaining to make ChatGPT-4o generate texts longer than usual. After succeeding, they created a training dataset with longer texts and fine-tuned two current models, Llama 3 and GLM 4. I tried the GLM 4 on Hugging Face and fed it a brief book summary of Jonathan Safran Foer's "Eating Animals." The result is ten pages of text, approximately 30,000 characters. The, well, good news: The result isn't a copy of Foer. It's AI slop, barely readable at best. Which is also the bad news, I guess.

30

7 Kommentare

Last month, DeepMind released FACTS Grounding - a benchmark for evaluating Factuality and Quality of LLM outputs...and they avoided most of the common benchmark pitfalls that I check for! (https://lnkd.in/eT88wNHj) - ✅ *Reproducibility*: Official evaluation code is available (https://lnkd.in/exp3kP2c) - ✅ *Test Data Leakage*: About half of the evaluation examples are private - they can not accidentally be scraped or included in an instruction tuning dataset (unlike GSM8K - https://lnkd.in/ec_a87ZJ) - ✅ *Construct Validity*: The benchmark design involves a step to check if the output is "eligible" (i.e. that it contains the relevant information) - this discredits the responses that are "factual" but contain no useful information. The developers, also, did not include examples that involved creativity, mathematics, or complex reasoning to target factuality exclusively. - ✅ *Saturation*: Over time SotA Models hit a ceiling on performance, and no longer represent a useful proxy for performance. The top models currently score in the low 80s on this benchmark, which shows that there is still room for improvement. - ✅ *Process and Limitation Documentation*: The benchmark developers clearly document the process for assembling the benchmark and explain how they handled limitations. For example: Evaluation is done using LLM-as-a-Judge, and "judges" are known to favor outputs from the same model. The evaluation uses an ensemble of 3 different models to reduce this bias. - 🚫 *Variance*: Each input prompt is only run once for every model. Given that LLMs produce different outputs over multiple runs of the same input, I would have liked to see that element incorporated in evaluation. However, given that LLM-as-a-Judge methods are used during grading, the one-run outputs may be fairly reliable. I am still finalizing my checklist of common benchmark pitfalls, so stay tuned.

19

https://lnkd.in/ef_gzFdU Time series data is super cool. Your outputs from a future time are your labels for the present time. This particular tutorial is especially good as it teaches many aspects of software/ML engineering. However, in my opinion, there are some shortcomings with the approach outlined. Let's see how many we can spot. For the googlers in my connections, if anyone knows how we can submit code to improve upon this, would love to know how.

13

2 full-length tutorial videos showing you how to implement various agentic RAG techniques - adding LLM layers to reason over inputs and post process the outputs.

1

Ensemble of smaller LLMs (PoLL) is less biased, faster, 7x cheaper when evaluating LLMs performance. Works great on QA & Arena-hard evals arxiv.org/abs/2404.18796

9