---
base_model: allenai/digital-socrates-13b
inference: false
language: en
library_name: transformers
license: apache-2.0
model_creator: Allen Institute for AI
model_name: Digital Socrates 13B
model_type: llama
prompt_template: '[INST] <>
{system_message}
<>
{prompt} [/INST]
'
quantized_by: TheBloke
---
# Digital Socrates 13B - GGUF
- Model creator: [Allen Institute for AI](https://huggingface.co/allenai)
- Original model: [Digital Socrates 13B](https://huggingface.co/allenai/digital-socrates-13b)
## Description
This repo contains GGUF format model files for [Allen Institute for AI's Digital Socrates 13B](https://huggingface.co/allenai/digital-socrates-13b).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/digital-socrates-13B-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/digital-socrates-13B-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF)
* [Allen Institute for AI's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/allenai/digital-socrates-13b)
## Prompt template: Llama-2-Chat
```
[INST] <>
{system_message}
<>
{prompt} [/INST]
```
## Licensing
The creator of the source model has listed its license as `apache-2.0`, and this quantization has therefore used that same license.
As this model is based on Llama 2, it is also subject to the Meta Llama 2 license terms, and the license files for that are additionally included. It should therefore be considered as being claimed to be licensed under both licenses. I contacted Hugging Face for clarification on dual licensing but they do not yet have an official position. Should this change, or should Meta provide any feedback on this situation, I will update this section accordingly.
In the meantime, any questions regarding licensing, and in particular how these two licenses might interact, should be directed to the original model repository: [Allen Institute for AI's Digital Socrates 13B](https://huggingface.co/allenai/digital-socrates-13b).
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
Click to see details
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [digital-socrates-13b.Q2_K.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q2_K.gguf) | Q2_K | 2 | 5.43 GB| 7.93 GB | smallest, significant quality loss - not recommended for most purposes |
| [digital-socrates-13b.Q3_K_S.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q3_K_S.gguf) | Q3_K_S | 3 | 5.66 GB| 8.16 GB | very small, high quality loss |
| [digital-socrates-13b.Q3_K_M.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q3_K_M.gguf) | Q3_K_M | 3 | 6.34 GB| 8.84 GB | very small, high quality loss |
| [digital-socrates-13b.Q3_K_L.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q3_K_L.gguf) | Q3_K_L | 3 | 6.93 GB| 9.43 GB | small, substantial quality loss |
| [digital-socrates-13b.Q4_0.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q4_0.gguf) | Q4_0 | 4 | 7.37 GB| 9.87 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [digital-socrates-13b.Q4_K_S.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q4_K_S.gguf) | Q4_K_S | 4 | 7.41 GB| 9.91 GB | small, greater quality loss |
| [digital-socrates-13b.Q4_K_M.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q4_K_M.gguf) | Q4_K_M | 4 | 7.87 GB| 10.37 GB | medium, balanced quality - recommended |
| [digital-socrates-13b.Q5_0.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q5_0.gguf) | Q5_0 | 5 | 8.97 GB| 11.47 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [digital-socrates-13b.Q5_K_S.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q5_K_S.gguf) | Q5_K_S | 5 | 8.97 GB| 11.47 GB | large, low quality loss - recommended |
| [digital-socrates-13b.Q5_K_M.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q5_K_M.gguf) | Q5_K_M | 5 | 9.23 GB| 11.73 GB | large, very low quality loss - recommended |
| [digital-socrates-13b.Q6_K.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q6_K.gguf) | Q6_K | 6 | 10.68 GB| 13.18 GB | very large, extremely low quality loss |
| [digital-socrates-13b.Q8_0.gguf](https://huggingface.co/TheBloke/digital-socrates-13B-GGUF/blob/main/digital-socrates-13b.Q8_0.gguf) | Q8_0 | 8 | 13.83 GB| 16.33 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/digital-socrates-13B-GGUF and below it, a specific filename to download, such as: digital-socrates-13b.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/digital-socrates-13B-GGUF digital-socrates-13b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
More advanced huggingface-cli download usage
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/digital-socrates-13B-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/digital-socrates-13B-GGUF digital-socrates-13b.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m digital-socrates-13b.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "[INST] <>\n{system_message}\n<>\n{prompt} [/INST]"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p ` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/digital-socrates-13B-GGUF", model_file="digital-socrates-13b.Q4_K_M.gguf", model_type="llama", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Brandon Frisco, LangChain4j, Spiking Neurons AB, transmissions 11, Joseph William Delisle, Nitin Borwankar, Willem Michiel, Michael Dempsey, vamX, Jeffrey Morgan, zynix, jjj, Omer Bin Jawed, Sean Connelly, jinyuan sun, Jeromy Smith, Shadi, Pawan Osman, Chadd, Elijah Stavena, Illia Dulskyi, Sebastain Graf, Stephen Murray, terasurfer, Edmond Seymore, Celu Ramasamy, Mandus, Alex, biorpg, Ajan Kanaga, Clay Pascal, Raven Klaugh, 阿明, K, ya boyyy, usrbinkat, Alicia Loh, John Villwock, ReadyPlayerEmma, Chris Smitley, Cap'n Zoog, fincy, GodLy, S_X, sidney chen, Cory Kujawski, OG, Mano Prime, AzureBlack, Pieter, Kalila, Spencer Kim, Tom X Nguyen, Stanislav Ovsiannikov, Michael Levine, Andrey, Trailburnt, Vadim, Enrico Ros, Talal Aujan, Brandon Phillips, Jack West, Eugene Pentland, Michael Davis, Will Dee, webtim, Jonathan Leane, Alps Aficionado, Rooh Singh, Tiffany J. Kim, theTransient, Luke @flexchar, Elle, Caitlyn Gatomon, Ari Malik, subjectnull, Johann-Peter Hartmann, Trenton Dambrowitz, Imad Khwaja, Asp the Wyvern, Emad Mostaque, Rainer Wilmers, Alexandros Triantafyllidis, Nicholas, Pedro Madruga, SuperWojo, Harry Royden McLaughlin, James Bentley, Olakabola, David Ziegler, Ai Maven, Jeff Scroggin, Nikolai Manek, Deo Leter, Matthew Berman, Fen Risland, Ken Nordquist, Manuel Alberto Morcote, Luke Pendergrass, TL, Fred von Graf, Randy H, Dan Guido, NimbleBox.ai, Vitor Caleffi, Gabriel Tamborski, knownsqashed, Lone Striker, Erik Bjäreholt, John Detwiler, Leonard Tan, Iucharbius
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
# Original model card: Allen Institute for AI's Digital Socrates 13B
This is the Digital Socrates 13B (DS-13B) model described in our paper: Digital Socrates: Evaluating LLMs through explanation critiques (arXiv link: https://arxiv.org/abs/2311.09613).
The less recommended, smaller 7B model can be found at https://huggingface.co/allenai/digital-socrates-7b
The DS-13B model is a fine-tuned version of [Llama-2-13b-Chat](https://huggingface.co/meta-llama/Llama-2-13b-chat), please
review its [guidelines](https://huggingface.co/meta-llama/Llama-2-13b-chat/blob/main/USE_POLICY.md) and
[licensing](https://huggingface.co/meta-llama/Llama-2-13b-chat#model-details) before use.
# What is Digital Socrates?
Digital Socrates is an open-source, automatic explanation-critiquing model. It is useful for revealing insights about student
models by examining their reasoning chains. The use of Digital Socrates allows for nuanced, interpretable automatic
evaluation of explanations without expensive API calls or human annotations.
While LLMs can provide reasoned explanations along with their answers, the nature and
quality of those explanations are still poorly understood. In response, our goal is to define a detailed way of characterizing the
explanation capabilities of modern models and to create a nuanced, interpretable explanation
evaluation tool that can generate such characterizations automatically.
Given a **question** (together with the multiple-choice options and gold answer),
```commandline
When Dennis operates his lawnmower, he notices the engine makes a lot of noise. He also notices that the engine gets very hot. Which best describes the heat and noise generated from the lawnmower?
(A) a change in phase
(B) thermal expansion
(C) an increase in entropy
(D) mechanical advantage
Answer sheet: (C)
```
as well as a student model explanation and answer,
```
1) The question states that the lawnmower engine makes a lot of noise.
2) The question states that the lawnmower engine gets very hot.
3) Noise and heat are both forms of energy.
4) The noise and heat generated from the lawnmower are a result of the conversion of energy from the fuel to mechanical energy.
Answer: (D)
```
Digital Socrates gives a **critique of the model-generated explanation** that provides localized feedback on the
most significant flaw (if any) in the explanation, suggestions to address the identified flaw, as well as a numeric rating indicating the explanation quality:
```
The explanation states or suggests the following:
* Main flaw (standalone statement): "The noise and heat generated from the lawnmower are a result of the conversion of energy from the fuel to mechanical energy."
* Dimension: incorrect_information
Consider these points for revising the explanation:
* General: It's important to understand the difference between the different types of energy. Mechanical energy is the energy of motion, while thermal energy is the energy of heat.
* Specific: In the case of the lawnmower, the noise and heat are not a result of the conversion of energy from the fuel to mechanical energy. The noise is a result of the vibration of the engine, while the heat is a result of the friction and combustion of the fuel.
Explanation score: 2
```
Remarkably, despite being orders of magnitude smaller than GPT-4, our Digital Socrates models are
capable of generating critiques close to GPT-4 critiques in terms of human rating and other
quantitative measures (correlation of explanation scores given and error category matches).
Through quantitative and qualitative analysis, we demonstrate how Digital Socrates is useful for
revealing insights about student models by examining their reasoning chains.
We invite you to try out Digital Socrates for your own application!
# How to use Digital Socrates?
We provide a quick example of how you can try out Digital Socrates with just a few lines of code:
'DSCritiqueBank-V1' used below can be downloaded from our [dataset page](https://allenai.org/data/digital-socrates).
```
import json
from transformers import AutoTokenizer, AutoModelForCausalLM
# Load model and tokenizer
model_path = "allenai/digital-socrates-13b"
model = AutoModelForCausalLM.from_pretrained(model_path, device_map="auto")
tokenizer = AutoTokenizer.from_pretrained(model_path)
# Define input data
question = "When Dennis operates his lawnmower, he notices the engine makes a lot of noise. He also notices that the engine gets very hot. Which best describes the heat and noise generated from the lawnmower? (A) a change in phase (B) thermal expansion (C) an increase in entropy (D) mechanical advantage"
explanation = "1) The question states that the lawnmower engine makes a lot of noise.\n2) The question states that the lawnmower engine gets very hot.\n3) Noise and heat are both forms of energy.\n4) The noise and heat generated from the lawnmower are a result of the conversion of energy from the fuel to mechanical energy."
answerkey = "C"
predictedanswer = "D"
# construct prompt (Llama conventions)
with open("../DSCritiqueBank-V1/DSCB-prompts.json") as file:
prompts = json.load(file)
system_prompt = prompts['digital_socrates_v1']['system']
user_prompt = prompts['digital_socrates_v1']['main'].replace("[[QUESTION]]", question).replace("[[EXPLANATION]]", explanation).replace("[[PREDICTEDANSWER]]", predictedanswer).replace("[[ANSWERKEY]]", answerkey)
full_prompt = f"[INST] <>\n{system_prompt}\n<{user_prompt} [/INST]\n\n"
# Run model
input_ids = tokenizer.encode(full_prompt, return_tensors="pt").to("cuda:0")
output = model.generate(input_ids, max_new_tokens=512, temperature=0)
res = tokenizer.batch_decode(output, skip_special_tokens=True)
```
Print the output:
```
>>> print(res[0].split("[/INST]")[-1])
The explanation states or suggests the following:
* Main flaw (standalone statement): "The noise and heat generated from the lawnmower are a result of the conversion of energy from the fuel to mechanical energy."
* Dimension: incorrect_information
Consider these points for revising the explanation:
* General: It's important to understand the difference between the different types of energy. Mechanical energy is the energy of motion, while thermal energy is the energy of heat.
* Specific: In the case of the lawnmower, the noise and heat are not a result of the conversion of energy from the fuel to mechanical energy. The noise is a result of the vibration of the engine, while the heat is a result of the friction and combustion of the fuel.
Explanation score: 2
```
# More details about Digital Socrates ...
For more details about Digital Socrates, please refer to our:
* 📄Paper: https://arxiv.org/abs/2311.09613
* 💻Dataset: https://allenai.org/data/digital-socrates
