context-cite

Attribute (or cite) statements generated by LLMs back to in-context information.

Stars: 110

Visit

ContextCite is a tool for attributing statements generated by LLMs back to specific parts of the context. It allows users to analyze and understand the sources of information used by language models in generating responses. By providing attributions, users can gain insights into how the model makes decisions and where the information comes from.

README:

ContextCite: Attributing Model Generation to Context

[getting started] [example notebooks] [🤗 demo] [blog post #1] [blog post #2] [paper] [bib]
Maintainers: Ben Cohen-Wang, Harshay Shah, and Kristian Georgiev

context_cite is a tool for attributing statements generated by LLMs back to specific parts of the context.

Getting started

Install context_cite via pip

pip install context_cite

Using context_cite is as simple as:

from context_cite import ContextCiter

model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
context = """
Attention Is All You Need

Abstract
The dominant sequence transduction models are based on complex recurrent or convolutional neural networks that include an encoder and a decoder. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles, by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.
1 Introduction
Recurrent neural networks, long short-term memory [13] and gated recurrent [7] neural networks in particular, have been firmly established as state of the art approaches in sequence modeling and transduction problems such as language modeling and machine translation [35, 2, 5]. Numerous efforts have since continued to push the boundaries of recurrent language models and encoder-decoder architectures [38, 24, 15].
Recurrent models typically factor computation along the symbol positions of the input and output sequences. Aligning the positions to steps in computation time, they generate a sequence of hidden states ht, as a function of the previous hidden state ht-1 and the input for position t. This inherently sequential nature precludes parallelization within training examples, which becomes critical at longer sequence lengths, as memory constraints limit batching across examples. Recent work has achieved significant improvements in computational efficiency through factorization tricks [21] and conditional computation [32], while also improving model performance in case of the latter. The fundamental constraint of sequential computation, however, remains.
Attention mechanisms have become an integral part of compelling sequence modeling and transduction models in various tasks, allowing modeling of dependencies without regard to their distance in the input or output sequences [2, 19]. In all but a few cases [27], however, such attention mechanisms are used in conjunction with a recurrent network.
In this work we propose the Transformer, a model architecture eschewing recurrence and instead relying entirely on an attention mechanism to draw global dependencies between input and output. The Transformer allows for significantly more parallelization and can reach a new state of the art in translation quality after being trained for as little as twelve hours on eight P100 GPUs.
"""
query = "What type of GPUs did the authors use in this paper?"
cc = ContextCiter.from_pretrained(model_name, context, query, device="cuda")

We can check the model's response using cc.response:

In [1]: cc.response
Out[1]: 'The authors used eight P100 GPUs in their Transformer architecture for training on the WMT 2014 English-to-German translation task.</s>'

Where did the model get its information? Let's see what the attributions look like!

In [2]: cc.get_attributions(as_dataframe=True, top_k=5)
Out[2]:

Finally, let's try to attribute a specific part of the response. To do so, we specify a start_idx and end_idx corresponding to the range of the response that we would like to attribute. In this case, we'll specify indices to attribute the phrase "the WMT 2014 English-to-German translation task" from the response.

In [3]: cc.get_attributions(start_idx=83, end_idx=129, as_dataframe=True, top_k=5)
Out[3]:

Example notebooks

Try out context_cite using our example notebooks (you can open them in Google colab):

🤗 quickstart: a quick introduction to context_cite
🦜⛓️ RAG example: chaining context_cite within a simple langchain RAG setup

Citation

@misc{cohenwang2024contextciteattributingmodelgeneration,
      title={ContextCite: Attributing Model Generation to Context}, 
      author={Benjamin Cohen-Wang and Harshay Shah and Kristian Georgiev and Aleksander Madry},
      year={2024},
      eprint={2409.00729},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2409.00729}, 
}

For Tasks:

Click tags to check more tools for each tasks

analyze responses understand sources attribute information interpret model decisions gain insights

For Jobs:

data scientist machine learning engineer research scientist ai ethicist nlp researcher

Alternative AI tools for context-cite

Similar Open Source Tools

context-cite

github

: 110

FigStep

FigStep is a black-box jailbreaking algorithm against large vision-language models (VLMs). It feeds harmful instructions through the image channel and uses benign text prompts to induce VLMs to output contents that violate common AI safety policies. The tool highlights the vulnerability of VLMs to jailbreaking attacks, emphasizing the need for safety alignments between visual and textual modalities.

github

: 52

BIG-Bench-Mistake

BIG-Bench Mistake is a dataset of chain-of-thought (CoT) outputs annotated with the location of the first logical mistake. It was released as part of a research paper focusing on benchmarking LLMs in terms of their mistake-finding ability. The dataset includes CoT traces for tasks like Word Sorting, Tracking Shuffled Objects, Logical Deduction, Multistep Arithmetic, and Dyck Languages. Human annotators were recruited to identify mistake steps in these tasks, with automated annotation for Dyck Languages. Each JSONL file contains input questions, steps in the chain of thoughts, model's answer, correct answer, and the index of the first logical mistake.

github

: 62

minbpe

This repository contains a minimal, clean code implementation of the Byte Pair Encoding (BPE) algorithm, commonly used in LLM tokenization. The BPE algorithm is "byte-level" because it runs on UTF-8 encoded strings. This algorithm was popularized for LLMs by the GPT-2 paper and the associated GPT-2 code release from OpenAI. Sennrich et al. 2015 is cited as the original reference for the use of BPE in NLP applications. Today, all modern LLMs (e.g. GPT, Llama, Mistral) use this algorithm to train their tokenizers. There are two Tokenizers in this repository, both of which can perform the 3 primary functions of a Tokenizer: 1) train the tokenizer vocabulary and merges on a given text, 2) encode from text to tokens, 3) decode from tokens to text. The files of the repo are as follows: 1. minbpe/base.py: Implements the `Tokenizer` class, which is the base class. It contains the `train`, `encode`, and `decode` stubs, save/load functionality, and there are also a few common utility functions. This class is not meant to be used directly, but rather to be inherited from. 2. minbpe/basic.py: Implements the `BasicTokenizer`, the simplest implementation of the BPE algorithm that runs directly on text. 3. minbpe/regex.py: Implements the `RegexTokenizer` that further splits the input text by a regex pattern, which is a preprocessing stage that splits up the input text by categories (think: letters, numbers, punctuation) before tokenization. This ensures that no merges will happen across category boundaries. This was introduced in the GPT-2 paper and continues to be in use as of GPT-4. This class also handles special tokens, if any. 4. minbpe/gpt4.py: Implements the `GPT4Tokenizer`. This class is a light wrapper around the `RegexTokenizer` (2, above) that exactly reproduces the tokenization of GPT-4 in the tiktoken library. The wrapping handles some details around recovering the exact merges in the tokenizer, and the handling of some unfortunate (and likely historical?) 1-byte token permutations. Finally, the script train.py trains the two major tokenizers on the input text tests/taylorswift.txt (this is the Wikipedia entry for her kek) and saves the vocab to disk for visualization. This script runs in about 25 seconds on my (M1) MacBook. All of the files above are very short and thoroughly commented, and also contain a usage example on the bottom of the file.

github

: 8.5k

GhostOS

GhostOS is an AI Agent framework designed to replace JSON Schema with a Turing-complete code interaction interface (Moss Protocol). It aims to create intelligent entities capable of continuous learning and growth through code generation and project management. The framework supports various capabilities such as turning Python files into web agents, real-time voice conversation, body movements control, and emotion expression. GhostOS is still in early experimental development and focuses on out-of-the-box capabilities for AI agents.

github

: 53

ChatAFL

ChatAFL is a protocol fuzzer guided by large language models (LLMs) that extracts machine-readable grammar for protocol mutation, increases message diversity, and breaks coverage plateaus. It integrates with ProfuzzBench for stateful fuzzing of network protocols, providing smooth integration. The artifact includes modified versions of AFLNet and ProfuzzBench, source code for ChatAFL with proposed strategies, and scripts for setup, execution, analysis, and cleanup. Users can analyze data, construct plots, examine LLM-generated grammars, enriched seeds, and state-stall responses, and reproduce results with downsized experiments. Customization options include modifying fuzzers, tuning parameters, adding new subjects, troubleshooting, and working on GPT-4. Limitations include interaction with OpenAI's Large Language Models and a hard limit of 150,000 tokens per minute.

github

: 257

LLMs-World-Models-for-Planning

This repository provides a Python implementation of a method that leverages pre-trained large language models to construct and utilize world models for model-based task planning. It includes scripts to generate domain models using natural language descriptions, correct domain models based on feedback, and support plan generation for tasks in different domains. The code has been refactored for better readability and includes tools for validating PDDL syntax and handling corrective feedback.

github

: 55

RAGElo

RAGElo is a streamlined toolkit for evaluating Retrieval Augmented Generation (RAG)-powered Large Language Models (LLMs) question answering agents using the Elo rating system. It simplifies the process of comparing different outputs from multiple prompt and pipeline variations to a 'gold standard' by allowing a powerful LLM to judge between pairs of answers and questions. RAGElo conducts tournament-style Elo ranking of LLM outputs, providing insights into the effectiveness of different settings.

github

: 98

llm-reasoners

LLM Reasoners is a library that enables LLMs to conduct complex reasoning, with advanced reasoning algorithms. It approaches multi-step reasoning as planning and searches for the optimal reasoning chain, which achieves the best balance of exploration vs exploitation with the idea of "World Model" and "Reward". Given any reasoning problem, simply define the reward function and an optional world model (explained below), and let LLM reasoners take care of the rest, including Reasoning Algorithms, Visualization, LLM calling, and more!

github

: 1.0k

llms

The 'llms' repository is a comprehensive guide on Large Language Models (LLMs), covering topics such as language modeling, applications of LLMs, statistical language modeling, neural language models, conditional language models, evaluation methods, transformer-based language models, practical LLMs like GPT and BERT, prompt engineering, fine-tuning LLMs, retrieval augmented generation, AI agents, and LLMs for computer vision. The repository provides detailed explanations, examples, and tools for working with LLMs.

github

: 266

monitors4codegen

This repository hosts the official code and data artifact for the paper 'Monitor-Guided Decoding of Code LMs with Static Analysis of Repository Context'. It introduces Monitor-Guided Decoding (MGD) for code generation using Language Models, where a monitor uses static analysis to guide the decoding. The repository contains datasets, evaluation scripts, inference results, a language server client 'multilspy' for static analyses, and implementation of various monitors monitoring for different properties in 3 programming languages. The monitors guide Language Models to adhere to properties like valid identifier dereferences, correct number of arguments to method calls, typestate validity of method call sequences, and more.

github

: 179

LeanAide

LeanAide is a work in progress AI tool designed to assist with development using the Lean Theorem Prover. It currently offers a tool that translates natural language statements to Lean types, including theorem statements. The tool is based on GPT 3.5-turbo/GPT 4 and requires an OpenAI key for usage. Users can include LeanAide as a dependency in their projects to access the translation functionality.

github

: 68

neutone_sdk

The Neutone SDK is a tool designed for researchers to wrap their own audio models and run them in a DAW using the Neutone Plugin. It simplifies the process by allowing models to be built using PyTorch and minimal Python code, eliminating the need for extensive C++ knowledge. The SDK provides support for buffering inputs and outputs, sample rate conversion, and profiling tools for model performance testing. It also offers examples, notebooks, and a submission process for sharing models with the community.

github

: 468

multilspy

Multilspy is a Python library developed for research purposes to facilitate the creation of language server clients for querying and obtaining results of static analyses from various language servers. It simplifies the process by handling server setup, communication, and configuration parameters, providing a common interface for different languages. The library supports features like finding function/class definitions, callers, completions, hover information, and document symbols. It is designed to work with AI systems like Large Language Models (LLMs) for tasks such as Monitor-Guided Decoding to ensure code generation correctness and boost compilability.

github

: 175

forust

Forust is a lightweight package for building gradient boosted decision tree ensembles. The algorithm code is written in Rust with a Python wrapper. It implements the same algorithm as XGBoost and provides nearly identical results. The package was developed to better understand XGBoost, as a fun project in Rust, and to experiment with adding new features to the algorithm in a simpler codebase. Forust allows training gradient boosted decision tree ensembles with multiple objective functions, predicting on datasets, inspecting model structures, calculating feature importance, and saving/loading trained boosters.

github

: 53

BentoDiffusion

BentoDiffusion is a BentoML example project that demonstrates how to serve and deploy diffusion models in the Stable Diffusion (SD) family. These models are specialized in generating and manipulating images based on text prompts. The project provides a guide on using SDXL Turbo as an example, along with instructions on prerequisites, installing dependencies, running the BentoML service, and deploying to BentoCloud. Users can interact with the deployed service using Swagger UI or other methods. Additionally, the project offers the option to choose from various diffusion models available in the repository for deployment.

github

: 325

For similar tasks

lmql

LMQL is a programming language designed for large language models (LLMs) that offers a unique way of integrating traditional programming with LLM interaction. It allows users to write programs that combine algorithmic logic with LLM calls, enabling model reasoning capabilities within the context of the program. LMQL provides features such as Python syntax integration, rich control-flow options, advanced decoding techniques, powerful constraints via logit masking, runtime optimization, sync and async API support, multi-model compatibility, and extensive applications like JSON decoding and interactive chat interfaces. The tool also offers library integration, flexible tooling, and output streaming options for easy model output handling.

github

: 3.4k

context-cite

github

: 110

GenAiGuidebook

GenAiGuidebook is a comprehensive resource for individuals looking to begin their journey in GenAI. It serves as a detailed guide providing insights, tips, and information on various aspects of GenAI technology. The guidebook covers a wide range of topics, including introductory concepts, practical applications, and best practices in the field of GenAI. Whether you are a beginner or an experienced professional, this resource aims to enhance your understanding and proficiency in GenAI.

github

: 51

ExplainableAI.jl

ExplainableAI.jl is a Julia package that implements interpretability methods for black-box classifiers, focusing on local explanations and attribution maps in input space. The package requires models to be differentiable with Zygote.jl. It is similar to Captum and Zennit for PyTorch and iNNvestigate for Keras models. Users can analyze and visualize explanations for model predictions, with support for different XAI methods and customization. The package aims to provide transparency and insights into model decision-making processes, making it a valuable tool for understanding and validating machine learning models.

github

: 108

For similar jobs

prometheus-eval

Prometheus-Eval is a repository dedicated to evaluating large language models (LLMs) in generation tasks. It provides state-of-the-art language models like Prometheus 2 (7B & 8x7B) for assessing in pairwise ranking formats and achieving high correlation scores with benchmarks. The repository includes tools for training, evaluating, and using these models, along with scripts for fine-tuning on custom datasets. Prometheus aims to address issues like fairness, controllability, and affordability in evaluations by simulating human judgments and proprietary LM-based assessments.

github

: 581

cladder

CLadder is a repository containing the CLadder dataset for evaluating causal reasoning in language models. The dataset consists of yes/no questions in natural language that require statistical and causal inference to answer. It includes fields such as question_id, given_info, question, answer, reasoning, and metadata like query_type and rung. The dataset also provides prompts for evaluating language models and example questions with associated reasoning steps. Additionally, it offers dataset statistics, data variants, and code setup instructions for using the repository.

github

: 82

awesome-llm-unlearning

This repository tracks the latest research on machine unlearning in large language models (LLMs). It offers a comprehensive list of papers, datasets, and resources relevant to the topic.

github

: 149

COLD-Attack

COLD-Attack is a framework designed for controllable jailbreaks on large language models (LLMs). It formulates the controllable attack generation problem and utilizes the Energy-based Constrained Decoding with Langevin Dynamics (COLD) algorithm to automate the search of adversarial LLM attacks with control over fluency, stealthiness, sentiment, and left-right-coherence. The framework includes steps for energy function formulation, Langevin dynamics sampling, and decoding process to generate discrete text attacks. It offers diverse jailbreak scenarios such as fluent suffix attacks, paraphrase attacks, and attacks with left-right-coherence.

github

: 84

Awesome-LLM-in-Social-Science

Awesome-LLM-in-Social-Science is a repository that compiles papers evaluating Large Language Models (LLMs) from a social science perspective. It includes papers on evaluating, aligning, and simulating LLMs, as well as enhancing tools in social science research. The repository categorizes papers based on their focus on attitudes, opinions, values, personality, morality, and more. It aims to contribute to discussions on the potential and challenges of using LLMs in social science research.

github

: 333

awesome-llm-attributions

This repository focuses on unraveling the sources that large language models tap into for attribution or citation. It delves into the origins of facts, their utilization by the models, the efficacy of attribution methodologies, and challenges tied to ambiguous knowledge reservoirs, biases, and pitfalls of excessive attribution.

github

: 152

context-cite

github

: 110

confabulations

LLM Confabulation Leaderboard evaluates large language models based on confabulations and non-response rates to challenging questions. It includes carefully curated questions with no answers in provided texts, aiming to differentiate between various models. The benchmark combines confabulation and non-response rates for comprehensive ranking, offering insights into model performance and tendencies. Additional notes highlight the meticulous human verification process, challenges faced by LLMs in generating valid responses, and the use of temperature settings. Updates and other benchmarks are also mentioned, providing a holistic view of the evaluation landscape.

github

: 64