AI Safety

Tutorials

• Transformer Circuits Thread - Can we reverse engineer transformer language models into human-understandable computer programs? Interpretability research benefits a lot from interactive articles. As part of their effort, they've created several other resources besides their paper like "A Mathematical Framework for Transformer Circuits" and "toy models of superposition".
• Discovering Language Model Behaviors with Model-Written Evaluations (paper) - They automatically generate evaluations with LMs. They discover new cases of inverse scaling where LMs get worse with size. They also find some of the first examples of inverse scaling in RLHF, where more RLHF makes LMs worse.
• Transformers learn in-context by gradient descent (paper) - transformers-learn-in-context-by-gradient-descent)]
• Why Can GPT Learn In-Context? Language Models Secretly Perform Gradient Descent as Meta-Optimizers (paper)
• Cognitive Biases in Large Language Models
• Tracr: Compiled Transformers as a Laboratory for Interpretability (paper) - TRACR (TRAnsformer Compiler for RASP) is a compiler for converting RASP programs (DSL for Transformers) into weights of a GPT-like model. Usually, we train Transformers to encode algorithms in their weights. With TRACR, we go in the reverse direction; compile weights directly from explicit code. Why do this? Accelerate interpretability research. Think of it like formal methods (from software eng.) on Transformers. It can be difficult to check if the explanation an interpretability tool provides is correct. [Tweet, code]
• Yann LeCun's unwavering opinion on current (auto-regressive) LLMs (Tweet)
• Core Views on AI Safety: When, Why, What, and How
• Personalisation within bounds: A risk taxonomy and policy framework for the alignment of large language models with personalised feedback
• GPTs are GPTs: An Early Look at the Labor Market Impact Potential of Large Language Models (paper) - The paper argues GPT is a General Purpose Technology.
• Core Views on AI Safety: When, Why, What, and How

Articles

Additional Reading

• How to Build OpenAI's GPT-2: "The AI That's Too Dangerous to Release"
• OpenAI’s GPT2 - Food to Media hype or Wake Up Call?
• How the Transformers broke NLP leaderboards
• Real-time Natural Language Understanding with BERT using NVIDIA TensorRT
• NLP's Clever Hans Moment has Arrived
• Language, trees, and geometry in neural networks - a series of expository notes accompanying the paper, "Visualizing and Measuring the Geometry of BERT" by Google's People + AI Research (PAIR) team.
• Benchmarking Transformers: PyTorch and TensorFlow - a comparison of inference time (on CPU and GPU) and memory usage for a wide range of transformer architectures.
• Evolution of representations in the Transformer - An accessible article that presents the insights of their EMNLP 2019 paper. They look at how the representations of individual tokens in Transformers trained with different objectives change.
• The dark secrets of BERT - This post probes fine-tuned BERT models for linguistic knowledge. In particular, the authors analyse how many self-attention patterns with some linguistic interpretation are actually used to solve downstream tasks. TL;DR: They are unable to find evidence that linguistically interpretable self-attention maps are crucial for downstream performance.
• A Visual Guide to Using BERT for the First Time - Tutorial on using BERT in practice, such as for sentiment analysis on movie reviews by Jay Alammar.
• Turing-NLG: A 17-billion-parameter language model - us/research/blog/zero-deepspeed-new-system-optimizations-enable-training-models-with-over-100-billion-parameters).
• MUM (Multitask Unified Model): A new AI milestone for understanding information
• GPT-3 is No Longer the Only Game in Town - GPT-3 was by far the largest AI model of its kind last year (2020). Now? Not so much.
• OpenAI's API Now Available with No Waitlist - GPT-3 access without the wait. However, apps must be approved before going live. This release also allow them to review applications, monitor for misuse, and better understand the effects of this tech.
• The Inherent Limitations of GPT-3 - One thing missing from the article if you've read Gwern's GPT-3 Creative Fiction article before is the mystery known as "Repetition/Divergence Sampling":
• Building games and apps entirely through natural language using OpenAI's code-davinci model - The author built several small games and apps without touching a single line of code, simply by telling the model what they want.
• Open AI gets GPT-3 to work by hiring an army of humans to fix GPT’s bad answers
• GPT-3 can run code - You provide an input text and a command and GPT-3 will transform them into an expected output. It works well for tasks like changing coding style, translating between programming languages, refactoring, and adding doc. For example, converts JSON into YAML, translates Python code to JavaScript, improve the runtime complexity of the function.
• Using GPT-3 to explain how code works
• Character AI announces they're building a full stack AGI company - founders Noam Shazeer (co-invented Transformers, scaled them to supercomputers for the first time, and pioneered large-scale pretraining) and Daniel de Freitas (led the development of LaMDA), all of which are foundational to recent AI progress.
• How Much Better is OpenAI’s Newest GPT-3 Model? - In addition to ChatGPT, OpenAI releases text-davinci-003, a Reinforcement Learning-tuned model that performs better long-form writing. Example, it can explain code in the style of Eminem. 😀
• OpenAI rival Cohere launches language model API - Backed by AI experts, they aims to bring Google-quality predictive language to the masses. Aidan Gomez co-wrote a seminal 2017 paper at Google Brain that invented a concept known as "Transformers".
• Startups competing with OpenAI's GPT-3 all need to solve the same problems - Last year, two startups released their own proprietary text-generation APIs. AI21 Labs, launched its 178-billion-parameter Jurassic-1 in Aug 2021, and Cohere, released a range of models. Cohere hasn't disclosed how many parameters its models contain. ... There are other up-and-coming startups looking to solve the same issues. Anthropic, the AI safety and research company started by a group of ex-OpenAI employees. Several researchers have left Google Brain to join two new ventures started by their colleagues. One outfit is named Character.ai, and the other Persimmon Labs.
• Cohere Wants to Build the Definitive NLP Platform - Beyond generative models like GPT-3.
• State of AI Report 2022 - Key takeaways:
• The Scaling Hypothesis - On GPT-3: meta-learning, scaling, implications, and deep theory.
• AI And The Limits Of Language — An AI system trained on words and sentences alone will never approximate human understanding - What LLMs like ChatGPT can and cannot do, and why AGI is not here yet.
• The Next Generation Of Large Language Models - It highlights 3 emerging areas: 1) models that can generate their own training data to improve themselves, 2) models that can fact-check themselves, and 3) massive sparse expert models.
• GPT-4 analysis and predictions - Somehow related, in "Bing Chat is blatantly, aggressively misaligned" post, Gwern think how Bing Chat/Sydney can be so different from ChatGPT and his hypothesis: "Sydney is not a RLHF trained GPT-3 model but a GPT-4 model developed in a hurry". Some have also argued that Sydney performs better on reasoning tasks than ChatGPT/GPT-3.5 and it may be GPT-4.
• Mosaic LLMs (Part 2): GPT-3 quality for <$500k (2022) - They claimed their Composer PyTorch framework ease model training. Now with Colossal-AI framework, I wonder how good is their solution. Until their users train it, I guess everything is purely hypothetical.
• I made a transformer by hand (no training!) - Make a transformer to predict a simple sequence manually — not by training one, or using pretrained weights, but instead by assigning each weight, by hand, over an evening.
• OpenAI's API Now Available with No Waitlist - GPT-3 access without the wait. However, apps must be approved before going live. This release also allow them to review applications, monitor for misuse, and better understand the effects of this tech.
• Transformers From Scratch
• Finetune Llama 3.1 on GCP for production use cases
• Language Modelling at Scale: Gopher, Ethical considerations, and Retrieval - The paper present an analysis of Transformer-based language model performance across a wide range of model scales — from models with tens of millions of parameters up to a 280 billion parameter model called Gopher.
• Building games and apps entirely through natural language using OpenAI's code-davinci model - The author built several small games and apps without touching a single line of code, simply by telling the model what they want.
• Transformer Inference Arithmetic - up from Carol Chen, ML Ops at Cohere. This article presents detailed few-principles reasoning about LLM inference performance, with no experiments or difficult math.
• Startups competing with OpenAI's GPT-3 all need to solve the same problems - Last year, two startups released their own proprietary text-generation APIs. AI21 Labs, launched its 178-billion-parameter Jurassic-1 in Aug 2021, and Cohere, released a range of models. Cohere hasn't disclosed how many parameters its models contain. ... There are other up-and-coming startups looking to solve the same issues. Anthropic, the AI safety and research company started by a group of ex-OpenAI employees. Several researchers have left Google Brain to join two new ventures started by their colleagues. One outfit is named Character.ai, and the other Persimmon Labs.

Attention Concept

• Making Transformer networks simpler and more efficient - FAIR released an all-attention layer to simplify the Transformer model and an adaptive attention span method to make it more efficient (reduce computation time and memory footprint).

Attention Mechanism

• The Annotated Transformer by Harvard NLP Group - Further reading to understand the "Attention is all you need" paper.
• Attention? Attention! - Attention guide by Lilian Weng from OpenAI.
• Visualizing A Neural Machine Translation Model (Mechanics of Seq2seq Models With Attention)
• What Does BERT Look At? An Analysis of BERT’s Attention paper
• Fast Transformer Decoding: One Write-Head is All You Need (paper) - They proposed a variant of attention type called multi-query attention (MQA). The plain multi-head attention mechanism has one query, key, and value per head; multi-query instead shares one key and value across all of the different attention "heads". In practice, training time remains the same, but much faster to decode in inference. MQA significantly improves language models performance and efficiency. Users can get ~10x better throughput and ~30% lower latency on inference. However, MQA can lead to quality degradation, and moreover it may not be desirable to train a separate model just for faster inference. In 2022, PaLM, a decoder-style model and their use of MQA is an interesting architecture improvements over GPT. Recent models that use MQA include TII's Falcon (2023).
• GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints - They (1) propose a technique for uptraining existing multi-head attention (MHA) models into models with multi-query attention (MQA) using 5% of original pre-training compute, and (2) introduce grouped-query attention (GQA), a generalization of MQA which uses an intermediate (more than one, less than number of query heads) number of key-value heads. GQA achieves benefits close to MHA with comparable inference speed to MQA through reduced number of key-value heads. Models that use MQA include Meta's Llama 2 (2023). [Some Tweets]
• ![Visualizing Attention, a Transformer's Heart
• Neural Machine Translation by Jointly Learning to Align and Translate - Bahdanau invented the content-based neural attention that is now a core tool in deep-learning-based NLP (language models). A disadvantage of fixed-length context vector design is incapability of remembering long sentences. The attention mechanism was born to resolve this problem. It was born to help memorize long input sentences in language translation. [Bahdanau deserve the praise]
• Ring Attention with Blockwise Transformers for Near-Infinite Context - Ring Attention is a system-level optimization technique by leveraging specific hardware architecture to make the exact attention computation more efficient.
• Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention - Infini-attention has an additional compressive memory with linear attention for processing infinitely long contexts. They trained a 1B parameter Transformer model that was fine-tuned on up to 5K sequence length passkey instances solves the 1M tokens input length problem. The Infini-attention mechanism presents an efficient and powerful approach for Transformer language models to process very long contexts without prohibitive increases in memory or computation.
• Retrieval Head Mechanistically Explains Long-Context Factuality - The paper explains how LLMs actually deal with context windows. The findings: they discover LLMs have unexpectedly developed retrieval heads, they were not explicitly coded for by creators. [Code: An algorithm that statistically calculate the retrieval score of attention heads in a transformer model]
• Making Transformer networks simpler and more efficient - FAIR released an all-attention layer to simplify the Transformer model and an adaptive attention span method to make it more efficient (reduce computation time and memory footprint).

BERT and Transformer

• The Illustrated BERT, ELMo, and co. (How NLP Cracked Transfer Learning)
• Dissecting BERT - Understand BERT in depth with an intuitive, straightforward explanation of the relevant concepts.
• A Light Introduction to Transformer-XL
• Generalized Language Models
• ALBERT: A Lite BERT for Self-supervised Learning of Language Representations paper - layer parameter sharing, and Sentence Order Prediction (SOP) loss to model inter-sentence coherence. [Blog post | Code]
• ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators - Thang Luong, Quoc V. Le, and Christopher D. Manning - A BERT variant like ALBERT and cost less to train. They trained a model that outperforms GPT by using only one GPU; match the performance of RoBERTa by using 1/4 computation. It uses a new pre-training approach, called replaced token detection (RTD), that trains a bidirectional model while learning from all input positions. [Blog post | Code]
• Visual Paper Summary: ALBERT (A Lite BERT)
• Cramming: Training a Language Model on a Single GPU in One Day (paper) - While most in the community are asking how to push the limits of extreme computation, we ask the opposite question: How far can we get with a single GPU in just one day? ... Through the lens of scaling laws, we categorize a range of recent improvements to training and architecture and discuss their merit and practical applicability (or lack thereof) for the limited compute setting.
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• Open Sourcing BERT: State-of-the-Art Pre-training for Natural Language Processing
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What happened to BERT & T5? On Transformer Encoders, PrefixLM and Denoising Objectives
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT
• What is XLNet and why it outperforms BERT

Generative Pre-Training Transformer (GPT)

• Better Language Models and Their Implications
• Improving Language Understanding with Unsupervised Learning - this is an overview of the original OpenAI GPT model.
• 🦄 How to build a State-of-the-Art Conversational AI with Transfer Learning
• The Illustrated GPT-2 (Visualizing Transformer Language Models)
• MegatronLM: Training Billion+ Parameter Language Models Using GPU Model Parallelism
• OpenGPT-2: We Replicated GPT-2 Because You Can Too - the authors trained a 1.5 billion parameter GPT-2 model on a similar sized text dataset and they reported results that can be compared with the original model.
• MSBuild demo of an OpenAI generative text model generating Python code - The model that was trained on GitHub OSS repos. The model uses English-language code comments or simply function signatures to generate entire Python functions. Cool!
• GPT-3: Language Models are Few-Shot Learners (paper) - "We train GPT-3, an autoregressive language model with 175 billion parameters 😱, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting."
• How GPT3 Works - Visualizations and Animations
• GPT-4 Rumors From Silicon Valley - GPT-4 is almost ready. GPT-4 would be multimodal, accepting text, audio, image, and possibly video inputs. Release window: Dec - Feb. #hype
• New GPT-3 model: text-Davinci-003 - Improvements:
• GPT-4 research
• A Comprehensive Capability Analysis of GPT-3 and GPT-3.5 Series Models
• Sparks of Artificial General Intelligence: Early experiments with GPT-4 - There are completely mind-blowing examples in the paper.
• GPT-3.5 series - tuning by playing both sides of the conversation.
• People have noticed - 3 models.
• New GPT-3 model: text-Davinci-003 - Improvements:
• GPT-3.5 series - tuning by playing both sides of the conversation.
• Better Language Models and Their Implications
• What is ChatGPT?
• elyase/awesome-gpt3 - A collection of demos and articles about the OpenAI GPT-3 API.
• ChatGPT Universe
• GitHub Copilot - Codex is a descendant of GPT-3. Codex translates natural language into code.
• GitHub Copilot - Codex is a descendant of GPT-3. Codex translates natural language into code.
• GitHub Copilot - Codex is a descendant of GPT-3. Codex translates natural language into code.

Large Language Model (LLM)

• ChatGPT among the LLMs
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• Fun and Dystopia With AI-Based Code Generation Using GPT-J-6B - Prior to GitHub Copilot tech preview launch, Max Woolf, a data scientist tested GPT-J-6B's code "writing" abilities.
• GPT-NeoX-20B - A 20 billion parameter model trained using EleutherAI’s GPT-NeoX framework. They expect it to perform well on many tasks. You can try out the model on GooseAI playground.
• BigScience's BLOOM-176B - BLOOM is a 175-billion parameter model for language processing, able to generate text much like GPT-3 and OPT-175B. It was developed to be multilingual, being deliberately trained on datasets containing 46 natural languages and 13 programming languages.
• bitsandbytes-Int8 inference for Hugging Face models - You can run BLOOM-176B/OPT-175B easily on a single machine, without performance degradation. If true, this could be a game changer in enabling people outside of big tech companies being able to use these LLMs.
• WeLM: A Well-Read Pre-trained Language Model for Chinese (paper)
• Teaching Small Language Models to Reason (paper) - They finetune a student model on the chain of thought (CoT) outputs generated by a larger teacher model. For example, the accuracy of T5 XXL on GSM8K improves from 8.11% to 21.99% when finetuned on PaLM-540B generated chains of thought.
• ALERT: Adapting Language Models to Reasoning Tasks (paper) - They introduce ALERT, a benchmark and suite of analyses for assessing language models' reasoning ability comparing pre-trained and finetuned models on complex tasks that require reasoning skills to solve. It covers 10 different reasoning skills including logistic, causal, common-sense, abductive, spatial, analogical, argument and deductive reasoning as well as textual entailment, and mathematics.
• Evaluating Human-Language Model Interaction (paper) - They find that non-interactive performance does not always result in better human-LM interaction and that first-person and third-party metrics can diverge, suggesting the importance of examining the nuances of human-LM interaction.
• Unnatural Instructions: Tuning Language Models with (Almost) No Human Labor (paper) - instructions)] - Fine-tuning a T5 on a large dataset collected with virtually no human labor leads to a model that surpassing the performance of models such as T0++ and Tk-Instruct across various benchmarks. These results demonstrate the potential of model-generated data as a cost-effective alternative to crowdsourcing for dataset expansion and diversification.
• Rethinking with Retrieval: Faithful Large Language Model Inference (paper) - They shows the potential of enhancing LLMs by retrieving relevant external knowledge based on decomposed reasoning steps obtained through chain-of-thought (CoT) prompting. I predict we're going to see many of these types of retrieval-enhanced LLMs in 2023.
• REPLUG: Retrieval-Augmented Black-Box Language Models (paper) - TL;DR: Enhancing GPT-3 with world knowledge — a retrieval-augmented LM framework that combines a frozen LM with a frozen/tunable retriever. It improves GPT-3 in language modeling and downstream tasks by prepending retrieved documents to LM inputs. [Tweet]
• Progressive Prompts: Continual Learning for Language Models (paper) - Current LLMs have hard time with catastrophic forgetting and leveraging past experiences. The approach learns a prompt for new task and concatenates with frozen previously learned prompts. This efficiently transfers knowledge to future tasks. [code]
• Large Language Models Can Be Easily Distracted by Irrelevant Context (paper) - Adding the instruction "Feel free to ignore irrelevant information given in the questions." consistently improves robustness to irrelevant context.
• Toolformer: Language Models Can Teach Themselves to Use Tools (paper) - A smaller model trained to translate human intention into actions (i.e. decide which APIs to call, when to call them, what arguments to pass, and how to best incorporate the results into future token prediction).
• ERNIE 3.0 Titan: Exploring Larger-scale Knowledge Enhanced Pre-training for Language Understanding and Generation (paper) - ERNIE 3.0 Titan is the latest addition to Baidu's ERNIE (Enhanced Representation through kNowledge IntEgration) family. It's inspired by the masking strategy of Google's BERT. ERNIE is also a unified framework. They also proposed a controllable learning algorithm and a credible learning algorithm. They apply online distillation technique to compress their model. To their knowledge, it is the largest (260B parameters) Chinese dense pre-trained model so far. [article]
• Characterizing Attribution and Fluency Tradeoffs for Retrieval-Augmented Large Language Models (paper) - Despite recent progress, it has been difficult to prevent semantic hallucinations in generative LLMs. One common solution to this is augmenting LLMs with a retrieval system and making sure that the generated output is attributable to the retrieved information.
• Augmented Language Models (ALMs): a Survey (paper) - Augmenting language models with reasoning skills and the ability to use various, non-parametric external modules for context processing and outperform traditional LMs on several benchmarks. This new research direction has the potential to address interpretability, consistency and scalability issues.
• A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT (paper) - My remarks: this paper raises a lot of questions around the term "foundation models", i.e., what's the model bare minimum number of parameters to qualify as foundation? It sounds to me foundation models are an "invented" concept that doesn't have good validity.
• Multimodal Chain-of-Thought Reasoning in Language Models (paper) - The model outperform GPT-3.5 by 16% on the ScienceQA benchmark. This work is the first to study CoT reasoning in different modalities, language (text) and vision (images). Unfortunately, they never provide ablation study on how much of that performance gain was caused by the new modalities. [code]
• RECITE: Recitation-Augmented Language Models (paper) - How can ChatGPT-like models achieve greater factual accuracy without relying on an external retrieval search engine? This paper shows that recitation can help LLMs generate accurate factual knowledge by reciting relevant passages from their own memory (by sampling) before producing final answers. The core idea is motivated by the intuition: recite-step that recollects relevant knowledge pieces helps answer-step (generation) better output. That's a recite-answer paradigm: first ask the LLM to generate the support paragraphs that contain the answer (knowledge-recitation) and then use it as additional prompt, along with the question to ask the LLM to generate the answer. They verify the effectiveness on four LLMs. They also show that recitation can be more effective than retrieval. This is important since having a retriever may lead to unpredictable behavior (i.e., Bing/Sydney). [code]
• LLaMA: Open and Efficient Foundation Language Models (paper) - A collection of language models ranging from 7B to 65B parameters. LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla70B and PaLM-540B. This shows that smaller models trained with more data can outperform larger models. This is not contradictory to anything in the Chinchilla paper, because it's not compute-optimally trained. GPU hours for training 7B model=82,432, 65B model=1,022,362 😱. Total time spent for all models: 2048 A100-80GB GPU for a period of approximately 5 months. The 65B model cost something in the range of ~$1-4M. Access to the model will be granted on a case-by-case basis though. People interested can apply for access. (Mar 2: they just approved access to the models, llama-7B works in Colab cedrickchee/llama) [Takeaways: Tweet]
• Language Is Not All You Need: Aligning Perception with Language Models (paper) - They introduce KOSMOS-1, a Multimodal Large Language Model (MLLM) that can perceive general modalities, learn in context (i.e., few-shot), and follow instructions (i.e., zero-shot). The total number of parameters is about 1.6B.
• Check Your Facts and Try Again: Improving Large Language Models with External Knowledge and Automated Feedback - LLM-Augmenter significantly reduces ChatGPT's hallucinations without sacrificing the fluency and informativeness of its responses. The architecture and data flow: 1) Retrieve evidence from external knowledge. 2) Context and reasoning chains. 3) Give to LLM (i.e., ChatGPT). 4) Verify hallucinations. 5) If hallucinate, give feedback and revise.
• UL2: Unifying Language Learning Paradigms (paper) - UL2 is a unified framework for pretraining models that are universally effective across datasets and setups. Takeaways: Objective matters way more than architecture. Mixture-of-Denoisers (MoD) is effective if you care about doing well on more than one type of tasks/settings. UL2 frames different objective functions for training language models as denoising tasks, where the model has to recover missing sub-sequences of a given input. During pre-training it uses Mixture-of-Denoisers (MoD) that samples from a varied set of such objectives, each with different configurations. MoD combines diverse pre-training paradigms together. They demonstrated that models trained using the UL2 framework perform well in a variety of language domains, including prompt-based few-shot learning and models fine-tuned for down-stream tasks. They open sourced UL2 20B model and checkpoints back in 2022. In 2023, they open sourced Flan-UL2 20B and released the weights. Check out: [blog post, Tweet]. I'm excited to see what the community does with this new model.
• Larger language models do in-context learning (ICL) differently (paper) - Overriding semantic priors when presented with enough flipped labels is an emergent ability of scale. LLMs learn better mappings when ICL labels are semantically unrelated to inputs (i.e., apple/orange, negative/positive). Fine-tuning to follow instruction helps both. [Tweet]
• The BigScience ROOTS Corpus: A 1.6TB Composite Multilingual Dataset (paper) - Documents the data creation and curation efforts of Responsible Open-science Open-collaboration Text Source (ROOTS) corpus, a dataset used to train BLOOM. [Tweet]
• PanGu-Σ: Towards Trillion Parameter Language Model with Sparse Heterogeneous Computing (paper) - They develop a system that trained a trillion-parameter language model on a cluster of Ascend 910 AI processors and MindSpore framework. This resulted in a 6.3x increase in training throughput through heterogeneous computing.
• Context-faithful Prompting for Large Language Models (paper)
• Llama 2: Open Foundation and Fine-Tuned Chat Models (paper) - Llama 2 pretrained models are trained on 2 trillion tokens, and have double the context length than Llama 1. Its fine-tuned models have been trained on over 1 million human annotations. It outperforms other open source language models on many benchmarks. License: The model and weights are available for free for research and commercial use. It is not an open source model, rather an open approach model — for commercial use, your product cannot have more than 700 million monthly active users and requires a form to get access. Llama-2-chat is the new addition and is created through using supervised fine-tuning and then iteratively refined using RLHF. [Nathan Lambert's summary of the paper]
• Code Llama: Open Foundation Models for Code (paper) - Code Llama is a family of LLMs for code based on Llama 2 providing SoTA performance among open models, infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks. It is capable of generating code, and natural language about code, from both code and natural language prompts. It's available in three models: foundation models (Code Llama), Python specializations (Code Llama - Python), and instruction-following models (Code Llama - Instruct) with 7B, 13B and 34B parameters each. It outperformed publicly available LLMs benchmark on code tasks. They release it under the same permissive license (community license) as Llama 2.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• Introducing Llama 3: The most capable openly available LLM to date (article) - In the coming months, they’ll share the Llama 3 research paper. [Code]
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.
• Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context - The Gemini 1.5 model family technical report. Highlights: Gemini 1.5 Pro is now Google's most capable model (surpassing 1.0 Ultra), Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens.
• GPT-J-6B - Can't access GPT-3? Here's GPT-J — its open-source cousin.