Spaces:
Sleeping
Sleeping
File size: 4,276 Bytes
f229c82 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 |
import streamlit as st
st.set_page_config(page_title="HAERAE Open Research Questions", layout="wide")
st.title("HAERAE Open Research Questions")
st.write("""
HAERAE is a non-profit research lab focused on the interpretability and evaluation of Korean language models.
Our mission is to advance the field with insightful benchmarks and tools. Below is an overview of our projects.
We've been doing most of our projects internally, but for those that have been unsolvable,
we are planning to open them to get help from the open-source community.
""")
st.header("HAERAE-Math Challenge")
st.write("""
Today we are introducing our first challenge: HAERAE-Math. We've created high-quality instructions on math
but don't have an idea on how to generate high-quality answers for them. We are looking for solutions that
use open-source models with openly available licenses.
We have created a total of 20,000 instructions already and are generating more. We've opened up a preview
of 50 of them in this link: [HAERAE-Math Samples](https://huggingface.co/datasets/HAERAE-HUB/HAERAE-Math-samples)
For those who generate answers for the 50 and share the methodology/results with us, we'll share the
remaining instructions and credit for the resulting dataset.
""")
st.subheader("Example Question")
example_question = """
νκ΅μ 보μ μ λ¬Έκ°κ° κ³ λνλ λ°μ΄ν° λ³΄νΈ μμ€ν
μ κ°λ°νκ³ μμ΅λλ€. μ΄ μμ€ν
μ 3μ°¨μ κΈ°ννμ μ κΈ λ©μ»€λμ¦μ μ¬μ©νλλ°, μ κΈ μ₯μΉλ μλΏ λͺ¨μμΌλ‘ λμ΄ μκ³ , λ°λ©΄μ λ°μ§λ¦μ 6cm, λμ΄λ 8cmμ
λλ€. μ΄ μλΏ λͺ¨μμ μ κΈ μ₯μΉμλ μν΅ λͺ¨μμ μ΄μ κ° λ± λ§κ² λ€μ΄κ°κ² μ€κ³λμ΄ μμ΅λλ€.
보μ μ λ¬Έκ°λ λ λμ μμ€μ 보μμ μν΄ μν΅ λͺ¨μμ μ΄μ μμ ꡬ λͺ¨μμ μ κΈ μ₯μΉλ₯Ό μΆκ°νλ €κ³ ν©λλ€. μ΄ κ΅¬λ μν΅ μμ λ± λ€μ΄κ°λλ‘ μ€κ³λμ΄ μμ΅λλ€.
λ€μμ μ§λ¬Έλ€μ ν΄κ²°νμκΈ° λ°λλλ€:
1. μλΏ μμ λ± λ€μ΄κ°κ² μ€κ³λ μν΅μ λ°μ§λ¦μ μΌλ§μΈκ°μ?
2. μν΅ μμ λ± λ€μ΄κ°κ² μ€κ³λ ꡬμ λΆνΌλ μΌλ§μΈκ°μ?
3. μλΏ, μν΅, κ΅¬κ° λͺ¨λ κ°μ μ€μ¬μΆμ 곡μ νκ³ μμΌλ©° μλΏμ κΌλκΈ°μ κ³Ό μν΅, ꡬμ μ€μ¬μ μ΄ λμΌνλ€κ³ κ°μ νλ©΄, μλΏμμ μν΅μ΄ μ°¨μ§νλ λΉμ¨μ ꡬνμμ€.
4. μ΄μ μλΏμ λμ΄λ₯Ό 2λ°°λ‘ λ리μ. μλΏμ λμ΄κ° 16cmκ° λμμ λ, μν΅κ³Ό ꡬμ ν¬κΈ°μ λΆνΌλ μ΄λ»κ² λ³νλμ?
5. μλΏμ λμ΄μ λ°λ©΄μ λ°μ§λ¦μ κ°κ° hμ rμ΄λΌκ³ ν λ, μν΅κ³Ό ꡬμ μ΅λ λΆνΌλ₯Ό rκ³Ό hλ‘ νννμμ€.
μλΏ, μν΅, ꡬμ λΆνΌ 곡μμ μ¬μ©νμ¬ λ¬Έμ λ₯Ό ν΄κ²°νμκΈ° λ°λλλ€:
μλΏμ λΆνΌ: V = 1/3ΟrΒ²h
μν΅μ λΆνΌ: V = ΟrΒ²h
ꡬμ λΆνΌ: V = 4/3ΟrΒ³
"""
st.code(example_question, language="markdown")
st.header("How to Participate")
st.write("""
1. Access the 50 sample questions from the provided Hugging Face dataset link.
2. Generate high-quality answers for these questions using open-source models.
3. Document your methodology and results.
4. Share your findings with us through [contact information or submission form].
5. If your approach is promising, we'll provide access to the full dataset of 20,000 instructions.
6. Collaborate with us to refine and improve the answer generation process.
7. Receive credit as a contributor to the final HAERAE-Math dataset.
""")
st.header("Why Participate?")
st.write("""
- Contribute to advancing Korean language model research
- Gain access to a large, high-quality dataset of math instructions
- Collaborate with HAERAE researchers
- Receive recognition in the field of NLP and math education
- Potential for co-authorship on related publications
""")
st.header("Contact Us")
st.write("""
For more information or to submit your results, please contact us at:
[Your contact information or a link to a submission form]
""")
st.sidebar.title("About HAERAE")
st.sidebar.info("""
HAERAE is a non-profit research lab dedicated to advancing the field of
Korean language model interpretability and evaluation. Our work focuses on
creating insightful benchmarks and tools to push the boundaries of NLP research.
""") |