Commit History

  1. feat: Add initial documentation for smart contract problem and solution, including tasks for vulnerability detection and property discovery

    ajaxwin
    f93644a

  1. refactor: Update API base URL and model name, enhance message handling with history tracking

    ajaxwin
    e5b8b13
  2. refactor: Update task configurations and grading logic for improved scoring and consistency

    ajaxwin
    dccaaac
  3. fix: Update output instructions in prompts.py to enforce case sensitivity and structure, inference.py made DRY, PROJECT works

    ajaxwin
    c6002b4
  4. refactor: Update inference.py to use AsyncOpenAI and make episode functions asynchronous

    ajaxwin
    1248d28
  5. bug fixes in inference.py

    ajaxwin
    2171069
  6. refactor: Reward clamping in graders⁠

    ajaxwin
    41a051f
  7. refactor: Task3 reward model changed, agent adjusted for new model

    ajaxwin
    48661cd

  1. refactor: Improved grading logic for task 2

    ajaxwin
    f78cba2
  2. refactor: grader trivial bug,⁠

    ajaxwin
    7f7bcc6
  3. refactor: Update ActionType to include costs and modified grader for task 1

    ajaxwin
    5235476

  1. fix: Update environment variable loading to use load_dotenv and raise error for missing HF_TOKEN

    ajaxwin
    409c8b7
  2. refactor: Rename submit_function to submit and removes asserts from eval.py

    ajaxwin
    277ec6e
  3. Switching to old inference.py

    Ajaxwin
    Unverified signature
    8493010
  4. refactor: Update grading logic and submission handling across tasks for improved accuracy and consistency

    ajaxwin
    cfae7a7
  5. fix: Update environment variable loading to use load_dotenv and adjust episode count

    ajaxwin
    0b06e9e

  1. project working

    ajaxwin
    6be844b
  2. fix: Handle optional request body in reset endpoint and set default task ID

    ajaxwin
    a503619
  3. fix: Update API responses to return JSON format and remove deprecated file references

    ajaxwin
    cfd3cfa
  4. feat: Update inference script and Dockerfile for improved functionality⁠

    ajaxwin
    df6af9d
  5. fix: update file path for landing page in root endpoint

    ajaxwin
    7940abd
  6. html file path and api corrected

    ajaxwin
    c27e659
  7. fix: Update file paths and ensure model loading in PropertyRetriever

    ajaxwin
    45bd962
  8. feat: Add landing page and API documentation⁠

    ajaxwin
    17ed3a7

  1. fix import paths in app.py to reflect correct module structure

    ajaxwin
    0304fd3
  2. refactor Dockerfile: streamline build process by removing builder stage and consolidating dependency installation

    ajaxwin
    3c4240f
  3. refactor Dockerfile and requirements.txt: update base image, consolidate dependencies, and adjust NLTK data download

    ajaxwin
    144ccd7
  4. original dockerfile

    ajaxwin
    2031e4c
  5. refactor Dockerfile: consolidate package installation and remove packages.txt

    ajaxwin
    0225bde
  6. refactor Dockerfile: update CMD to ENTRYPOINT and remove packages.txt

    ajaxwin
    f21076f
  7. refactor Dockerfile: adjust NLTK downloader position and update venv path

    ajaxwin
    ed51cb8
  8. dockerfile nltk update

    ajaxwin
    1505111
  9. dockerfile updated again 3

    ajaxwin
    39dd0e3
  10. added packages.txt

    ajaxwin
    b1b6fba
  11. removed uv.lock

    ajaxwin
    b1a1c79
  12. dockerfile update 3

    ajaxwin
    a0858e7
  13. dockerfile updated

    ajaxwin
    878ca5a
  14. pyproject.toml updated

    ajaxwin
    7fe808d
  15. dockerfile, python version changed

    ajaxwin
    ac7fce1
  16. updated dockerfile

    ajaxwin
    fe328e6
  17. Readme -> Docs

    ajaxwin
    73c779f
  18. Structure Changed, files reviewed

    ajaxwin
    88875f7

  1. task3 evalutated

    ajaxwin
    1b91307
  2. key names updated

    ajaxwin
    2ee85c9
  3. task1, task2 evaluated⁠

    ajaxwin
    671787b

  1. task3 reviewed⁠

    ajaxwin
    056cf7b
  2. Task1 actions reviewed⁠

    ajaxwin
    c719864
  3. task2 reviewed, semanticmatcher implemented

    ajaxwin
    cf983b8

  1. separated actions, prompts

    ajaxwin
    e8c9acc
  2. Updated dataset with vulnerable functions

    ajaxwin
    49df6d3