Hugging Face's logo

Zero Shot NLI

non-profit
Activity Feed Request to join this org

AI & ML interests

None defined yet.

team-zero-shot-nli's activity

MoritzLaurer 
posted an update about 2 hours ago
view post
Post
94
Microsoft's rStar-Math paper claims that 🤏 ~7B models can match the math skills of o1 using clever train- and test-time techniques. You can now download their prompt templates from Hugging Face !

📏 The paper introduces rStar-Math, which claims to rival OpenAI o1's math reasoning capabilities by integrating Monte Carlo Tree Search (MCTS) with step-by-step verified reasoning trajectories.
🤖 A Process Preference Model (PPM) enables fine-grained evaluation of intermediate steps, improving training data quality.
🧪 The system underwent four rounds of self-evolution, progressively refining both the policy and reward models to tackle Olympiad-level math problems—without GPT-4-based data distillation.
💾 While we wait for the release of code and datasets, you can already download the prompts they used from the HF Hub!

Details and links here 👇
Prompt-templates docs: https://moritzlaurer.github.io/prompt_templates/
Templates on the hub: MoritzLaurer/rstar-math-prompts
Prompt-templates collection: MoritzLaurer/prompt-templates-6776aa0b0b8a923957920bb4
Paper: https://arxiv.org/pdf/2501.04519
MoritzLaurer 
posted an update 4 days ago
view post
Post
2824
FACTS is a great paper from @GoogleDeepMind on measuring the factuality of LLM outputs. You can now download their prompt templates from @huggingface to improve LLM-based fact-checking yourself!

📏 The paper introduces the FACTS Grounding benchmark for evaluating the factuality of LLM outputs.

🤖 Fact-checking is automated by an ensemble of LLM judges that verify if a response is fully grounded in a factual reference document.

🧪 The authors tested different prompt templates on held-out data to ensure their generalization.

📚 It's highly educational to read these templates to learn how frontier labs design prompts and understand their limitations.

💾 You can now download and reuse these prompt templates via the prompt-templates library!

🔄 The library simplifies sharing prompt templates on the HF hub or locally via standardized YAML files. Let’s make LLM work more transparent and reproducible by sharing more templates like this!

Links 👇
- prompt-templates docs: https://moritzlaurer.github.io/prompt_templates/
- all templates on the HF Hub: MoritzLaurer/facts-grounding-prompts
- FACTS paper: https://storage.googleapis.com/deepmind-media/FACTS/FACTS_grounding_paper.pdf
MoritzLaurer 
posted an update 6 days ago
view post
Post
1661
The TRL v0.13 release is 🔥! My highlight are the new process reward trainer to train models similar to o1 and tool call support:

🧠 Process reward trainer: Enables training of Process-supervised Reward Models (PRMs), which reward the quality of intermediate steps, promoting structured reasoning. Perfect for tasks like stepwise reasoning.

🔀 Model merging: A new callback leverages mergekit to merge models during training, improving performance by blending reference and policy models - optionally pushing merged models to the Hugging Face Hub.

🛠️ Tool call support: TRL preprocessing now supports tool integration, laying the groundwork for agent fine-tuning with examples like dynamic temperature fetching in prompts.

⚖️ Mixture of judges: The new AllTrueJudge combines decisions from multiple binary judges for more nuanced evaluation.

Read the release notes and other resources here 👇
Release: https://github.com/huggingface/trl/releases/tag/v0.13.0
Mergekit: https://github.com/arcee-ai/mergekit
Mixture of judges paper: The Perfect Blend: Redefining RLHF with Mixture of Judges (2409.20370)
MoritzLaurer 
posted an update 8 days ago
view post
Post
2029
OpenAI is losing money on the $200/month subscription 🤯. It's crazy how expensive it is to run these largest LLMs:

- ChatGPT Pro costs $200/month ($2,400/year) and is still unprofitable for OpenAI due to higher-than-expected usage.
- OpenAI reportedly expected losses of about $5 billion on revenue of $3.7 billion last year, with ChatGPT alone once costing an estimated $700,000 per day to operate. 💸🔥
- They build strong models and do great research. Whether this business model will work in the long run is one of the biggest questions in the AI economy today.

Source with the numbers 👇
https://techcrunch.com/2025/01/05/openai-is-losing-money-on-its-pricey-chatgpt-pro-plan-ceo-sam-altman-says/
·
MoritzLaurer 
posted an update 9 days ago
view post
Post
2178
🚀 Releasing a new zeroshot-classifier based on ModernBERT! Some key takeaways:

- ⚡ Speed & efficiency: It's multiple times faster and uses significantly less memory than DeBERTav3. You can use larger batch sizes and enabling bf16 (instead of fp16) gave me a ~2x speed boost as well
- 📉 Performance tradeoff: It performs slightly worse than DeBERTav3 on average across my zeroshot classification task collection
- 🧠 Use cases: I recommend using it for scenarios requiring speed and a larger context window (8k).
- 💡 What’s next? I’m preparing a newer version trained on better + longer synthetic data to fully leverage the 8k context window and improve upon the training mix of my older zeroshot-v2.0 models. I also hope that there will be a multilingual variant in the future.

Great work by https://huggingface.co/answerdotai !

If you’re looking for a high-speed zeroshot classifier, give it a try!

📄 Resources below: 👇
Base model: MoritzLaurer/ModernBERT-base-zeroshot-v2.0
Large model: MoritzLaurer/ModernBERT-large-zeroshot-v2.0
Updated zeroshot collection: MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f
ModernBERT collection with paper: answerdotai/modernbert-67627ad707a4acbf33c41deb
MoritzLaurer 
posted an update 26 days ago
view post
Post
2600
Quite excited by the ModernBERT release! 0.15/0.4B small, 2T modern pre-training data and tokenizer with code, 8k context window, great efficient model for embeddings & classification!

This will probably be the basis for many future SOTA encoders! And I can finally stop using DeBERTav3 from 2021 :D

Congrats @answerdotai , @LightOnIO and collaborators like @tomaarsen !

Paper and models here 👇https://huggingface.co/collections/answerdotai/modernbert-67627ad707a4acbf33c41deb
·
MoritzLaurer 
posted an update 29 days ago
MoritzLaurer 
posted an update about 1 month ago
view post
Post
1286
I've been building a small library for working with prompt templates on the HF hub: pip install prompt-templates. Motivation:

The community currently shares prompt templates in a wide variety of formats: in datasets, in model cards, as strings in .py files, as .txt/.yaml/.json/.jinja2 files etc. This makes sharing and working with prompt templates unnecessarily complicated.

Prompt templates are currently the main hyperparameter that people tune when building complex LLM systems or agents. If we don't have a common standard for sharing them, we cannot systematically test and improve our systems. After comparing different community approaches, I think that working with modular .yaml or .json files is the best approach.

The prompt-templates library :
- proposes a standard for sharing prompts (entirely locally or on the HF hub)
- provides some utilities that are interoperable with the broader ecosystem

Try it: 
# !pip install prompt-templates
from prompt_templates import PromptTemplateLoader 
prompt_template = PromptTemplateLoader.from_hub(repo_id="MoritzLaurer/closed_system_prompts", filename="claude-3-5-artifacts-leak-210624.yaml")


The library is in early stages, feedback is welcome!

More details in the docs: https://github.com/MoritzLaurer/prompt_templates/
  • 1 reply
·
MoritzLaurer 
posted an update 3 months ago
view post
Post
4577
#phdone - I defended my PhD yesterday! A key lesson: it is amazing how open science and open source can empower beginners with limited resources:

I first learned about instruction-based classifiers like BERT-NLI 3-4 years ago, through the @HuggingFace ZeroShotClassificationPipeline. Digging deeper into this, it was surprisingly easy to find new datasets, newer base models, and reusable fine-tuning scripts on the HF Hub to create my own zeroshot models - although I didn't know much about fine-tuning at the time.

Thanks to the community effect of the Hub, my models were downloaded hundreds of thousands of times after a few months. Seeing my research being useful for people motivated me to improve and upload newer models. Leaving my contact details in the model cards led to academic cooperation and consulting contracts (and eventually my job at HF).

That's the power of open science & open source: learning, sharing, improving, collaborating.

I mean every word in my thesis acknowledgments (screenshot). I'm very grateful to my supervisors @vanatteveldt @CasAndreu @KasperWelbers for their guidance; to @profAndreaRenda and @CEPS_thinktank for enabling me to work part-time during the first year; to @huggingface for creating awesome tools and an awesome platform; and to many others who are not active on social media.

Links to the full thesis and the collection of my most recent models are below.

PS: If someone happens to speak Latin, let me know if my diploma contains some hidden Illuminati code or something :D
·
MoritzLaurer 
posted an update 4 months ago
view post
Post
2305
The new NIM Serverless API by HF and Nvidia is a great option if you want a reliable API for open-weight LLMs like Llama-3.1-405B that are too expensive to run on your own hardware.

- It's pay-as-you-go, so it doesn't have rate limits like the standard HF Serverless API and you don't need to commit to hardware like for a dedicated endpoint.
- It works out-of-the box with the new v0.25 release of our huggingface_hub.InferenceClient
- It's specifically tailored to a small collection of popular open-weight models. For a broader selection of open models, we recommend using the standard HF Serverless API.
- Note that you need a token from an Enterprise Hub organization to use it.

Details in this blog post: https://huggingface.co/blog/inference-dgx-cloud
Compatible models in this HF collection: nvidia/nim-serverless-inference-api-66a3c6fcdcb5bbc6e975b508
Release notes with many more features of huggingface_hub==0.25.0: https://github.com/huggingface/huggingface_hub/releases/tag/v0.25.0

Copy-pasteable code in the first comment:
·
MoritzLaurer 
posted an update 4 months ago
view post
Post
1628
Why would you fine-tune a model if you can just prompt an LLM? The new paper "What is the Role of Small Models in the LLM Era: A Survey" provides a nice pro/con overview. My go-to approach combines both:

1. Start testing an idea by prompting an LLM/VLM behind an API. It's fast and easy and I avoid wasting time on tuning a model on a task that might not make it into production anyways.

2. The LLM/VLM then needs to be manually validated. Anyone seriously considering putting AI into production has to do at least some manual validation. Setting up a good validation pipeline with a tool like Argilla is crucial and it can be reused for any future experiments. Note: you can use LLM-as-a-judge to automate some evals, but you always also need to validate the judge!

3. Based on this validation I can then (a) either just continue using the prompted LLM if it is accurate enough and it makes sense financially given my load; or (b) if the LLM is not accurate enough or too expensive to run in the long-run, I reuse the existing validation pipeline to annotate some additional data for fine-tuning a smaller model. This can be sped up by reusing & correcting synthetic data from the LLM (or just pure distillation).

Paper: https://arxiv.org/pdf/2409.06857
Argilla docs: https://docs.argilla.io/latest/
Argilla is also very easy to deploy with Hugging Face Spaces (or locally): https://huggingface.co/new-space?template=argilla%2Fargilla-template-space
MoritzLaurer 
posted an update 8 months ago
view post
Post
3304
We are hiring a "Developer Experience Engineer for Inference" at Hugging Face! If you want to make it easier for millions of people to use modern machine learning inference, apply! You can either work from one of our offices e.g. in Paris or New York, or work fully remotely. Details: https://apply.workable.com/huggingface/j/E732F4B8FC/
MoritzLaurer 
posted an update 9 months ago
view post
Post
4732
Why does Meta invest millions in Llama 3 and then makes it available for free? Here is Zuckerberg's explanation to investors in the Q3 2023 earnings call:

"The second part of our playbook is open source software infrastructure. Our long-standing strategy has been to build and open source general infrastructure while keeping our specific product implementations proprietary.

[...] First, open source software is typically safer and more secure, as well as more compute efficient to operate due to all the ongoing feedback, scrutiny, and development from the community. This is a big deal because safety is one of the most important issues in AI. Efficiency improvements and lowering the compute costs also benefit everyone including us.

Second, open source software often becomes an industry standard, and when companies standardize on building with our stack, that then becomes easier to integrate new innovations into our products. That’s subtle, but the ability to learn and improve quickly is a huge advantage and being an industry standard enables that.

Third, open source is hugely popular with developers and researchers. We know that people want to work on open systems that will be widely adopted, so this helps us recruit the best people at Meta, which is a very big deal for leading in any new technology area.

And again, we typically have unique data and build unique product integrations anyway, so providing infrastructure like Llama as open source doesn't reduce our main advantages. This is why our long-standing strategy has been to open source general infrastructure and why I expect it to continue to be the right approach for us going forward."

Fully earnings call transcript: https://s21.q4cdn.com/399680738/files/doc_financials/2023/q4/META-Q4-2023-Earnings-Call-Transcript.pdf
MoritzLaurer 
posted an update 10 months ago
view post
Post
3709
🆕 Releasing a new series of 8 zeroshot classifiers: better performance, fully commercially useable thanks to synthetic data, up to 8192 tokens, run on any hardware.

Summary:
🤖 The zeroshot-v2.0-c series replaces commercially restrictive training data with synthetic data generated with mistralai/Mixtral-8x7B-Instruct-v0.1 (Apache 2.0). All models are released under the MIT license.
🦾 The best model performs 17%-points better across 28 tasks vs. facebook/bart-large-mnli (the most downloaded commercially-friendly baseline).
🌏 The series includes a multilingual variant fine-tuned from BAAI/bge-m3 for zeroshot classification in 100+ languages and with a context window of 8192 tokens
🪶 The models are 0.2 - 0.6 B parameters small, so they run on any hardware. The base-size models are +2x faster than bart-large-mnli while performing significantly better.
🤏 The models are not generative LLMs, they are efficient encoder-only models specialized in zeroshot classification through the universal NLI task.
🤑 For users where commercially restrictive training data is not an issue, I've also trained variants with even more human data for improved performance.

Next steps:
✍️ I'll publish a blog post with more details soon
🔮 There are several improvements I'm planning for v2.1. Especially the multilingual model has room for improvement.

All models are available for download in this Hugging Face collection: MoritzLaurer/zeroshot-classifiers-6548b4ff407bb19ff5c3ad6f

These models are an extension of the approach explained in this paper, but with additional synthetic data: https://arxiv.org/abs/2312.17543
·
HarrySaini 
authored a paper 11 months ago
khalidalt 
authored a paper 11 months ago
MoritzLaurer 
posted an update 11 months ago
view post
Post
Prompts are hyperparameters. Every time you test a different prompt on your data, you become less sure if the LLM actually generalizes to unseen data.

Issues of overfitting to a test set seem like concepts from boring times when people still fine-tuned models, but it's just as important for "zeroshot prompting". Using a separate validation split to tune the main hyperparameter of LLMs (the prompt) is just as important as train-val-test splitting for fine-tuning. The only difference is that you don't have a training dataset anymore and it somehow feels different because there is no training / no parameter updates.

Its easy to trick yourself into believing that an LLM performs well on your task, while you've actually overfit the prompt on your data. Every good "zeroshot" paper should clarify that they used a validation split for finding their prompt before final testing.
  • 1 reply
·
khalidalt 
authored a paper 11 months ago
khalidalt 
authored 2 papers 12 months ago