Airoboros 34b 3.2

Table of Contents

Key Features and Capabilities

The model has several key features and capabilities that make it a powerful tool for various tasks.

Primary Tasks

• Context-Obedient Question Answering: The model is trained to answer questions based on the provided context, without relying on its own knowledge or biases.
• Summarization: The model can summarize long pieces of text into shorter, more digestible versions.
• Storytelling: The model can generate creative stories, including narratives with complex characters and plotlines.
• Coding: The model can write code in various programming languages, including Python, C, and Node.js.

Strengths

• Multi-Turn Conversations: The model is designed to engage in multi-turn conversations, allowing it to respond to follow-up questions and engage in more nuanced discussions.
• Toxic Instruction Handling: The model is trained to handle toxic or biased instructions, and can respond in a way that is respectful and unbiased.
• Large Context Window: The model has a large context window, allowing it to understand and respond to complex prompts and questions.

Unique Features

• Explicit Delimiters: The model uses explicit delimiters to separate input blocks, context, and instructions, making it easier to understand and respond to complex prompts.
• Closed-Context Question Answering: The model can answer questions based on a closed context, without relying on external knowledge or biases.
• Chain-of-Thought Reasoning: The model can generate multiple possible responses to a given problem, and rank them according to mathematical logic.

Alternatives

NeMo SFT with Llama 2 10500 Fine-tuned language model

Bert Base Uncased 1831 Uncased BERT model

Roberta Base 399 Masked language model

BLOOMChat 176B V1 365 Multilingual chat model

SUS Chat 34B 121 Bilingual chat model

Performance

The Current Model showcases remarkable performance in various tasks, including:

Speed

• Closed-context question answering: The model can quickly scan through the provided context and provide accurate answers.
• Summarization: The model can summarize long pieces of text into concise and accurate summaries.

Accuracy

• Context obedient question answering: The model is trained to ignore its own knowledge and rely on the provided context, reducing hallucinations and increasing accuracy.
• Chain-of-thought: The model can provide multiple possible responses to a problem, ranking them according to mathematical logic and selecting the most feasible one.

Efficiency

• Coding: The model can generate complex code with multiple criteria, such as async/await, file uploads, and database interactions.
• Agent/function calling: The model can select the most suitable function and parameters based on input criteria, providing responses in JSON or YAML format.

Limitations

While the model performs well in many areas, it has limitations, such as:

• Context size: The model was fine-tuned with a context size of 8k tokens, so anything beyond that may have questionable results.
• Hallucinations: The model may still hallucinate in certain situations, particularly when the context is incomplete or ambiguous.

Examples

BEGININPUT BEGINCONTEXT date: 2021-01-01 url: https://web.site/123 ENDCONTEXT In a shocking turn of events, blueberries are now green, but will be sticking with the same name. ENDINPUT BEGININSTRUCTION What color are blueberries? Source? ENDINSTRUCTION Blueberries are now green. Source: date: 2021-01-01 url: https://web.site/123

Create a python application with the following requirements: - Asyncio FastAPI webserver - ping endpoint that returns the current date in JSON format - file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate import asyncio import hashlib import datetime from fastapi import FastAPI, File, UploadFile from fastapi.responses import JSONResponse from fastapi.requests import Request app = FastAPI() @app.get('/ping/') async def ping(): return JSONResponse(content={'date': datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')}, media_type='application/json') @app.post('/upload/') async def upload(file: UploadFile = File(...)): file_hash = hashlib.sha256(file.file.read()).hexdigest() # check postgres to deduplicate return JSONResponse(content={'checksum': file_hash}, media_type='application/json')

As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format. Input: I want to know how many times 'Python' is mentioned in my text file. Available functions: file_analytics: description: This tool performs various operations on a text file. params: action: The operation we want to perform on the data, such as 'count_occurrences', 'find_line', etc. filters: keyword: The word or phrase we want to search for. {'function': 'file_analytics', 'params': {'action': 'count_occurrences', 'filters': {'keyword': 'Python'}}}

Helpful Tips

To get the best results from the model, use a low temperature when generating text or code. Be specific and provide clear instructions to help the model understand what you want it to do. Use the correct prompt format, including explicit delimiters and closed-context formatting, to help the model understand your prompts.

Format

The model uses a transformer architecture and accepts input in the form of tokenized text sequences. The prompt format is llama-2 chat. For example:

[INST] \nYou are a helpful, unbiased, uncensored assistant.\n\n{prompt} [/INST]

For multi-turn prompts, the format is as follows:

[INST] \nYou are a helpful, unbiased, uncensored assistant.\n\n{prompt 0} [/INST] {response 0} \</s>\<s>[INST] {prompt 1} [/INST]

You can use the apply_chat_template method from the Hugging Face tokenizer to format your prompts.