Hugging Face's logo
Asankhaya Sharma's picture
43 23 144

Asankhaya Sharma

codelion
MrOvkill's profile picture Ah7med's profile picture AyoubChLin's profile picture
Β·
http://asankhaya.github.io/

AI & ML interests

AI/ML, Dev Tools and Application Security

Recent Activity

new activity 7 days ago
ShilongLiu/Grounding_DINO_demo:🚩 Report: Not working
new activity 7 days ago
codelion/Grounding_DINO_demo:Demo not working: Runtime error
updated a Space 7 days ago
codelion/Grounding_DINO_demo
View all activity

Organizations

meraGPT's profile picture Lambda Security's profile picture National University of Singapore's profile picture Patched's profile picture ZeroGPU Explorers's profile picture MLX Community's profile picture Social Post Explorers's profile picture

codelion's activity

reacted to melisa's post with ❀️πŸ”₯ 3 months ago
view post
Post
2998
πŸ”₯ Introducing "Writing in the Margins (WiM)" - better inference pattern for long context LLMs that solves the Lost-in-the-Middle problem πŸ”₯

Paper page: Writing in the Margins: Better Inference Pattern for Long Context Retrieval (2408.14906)

TL;DR
Make your model write "margin notes" as you chunk prefill the KV cache. Then ask it reread all notes before it speaks up.
Works with humans, works with AI πŸ€–

WiM leverages the chunked prefill of the key-value cache, which concurrently generates query-based extractive summaries at each step of the prefill that are subsequently reintegrated at the end of the computation. We term these intermediate outputs β€œmargins”, drawing inspiration from the practice of making margin notes for improved comprehension of long contexts in human reading. We show that this technique, which adds only minimal additional computation, significantly improves LLMs long context reasoning capabilities.

Think: Every chunk has a chance to be attended to/ be at the end of the context at least once. πŸŽ‰

πŸ“Š Results:
- An average accuracy boost of 7.5% in multi-hop reasoning tasks like HotpotQA and MultiHop-RAG.
- Even a 30% increase in F1-score for summarisation-like tasks (CWE).

Plus, WiM fits seamlessly into interactive applications (think: progress bar!). It can provide real-time progress updates during data retrieval and integration, making it user-friendly and transparent - a stark contrast to feeding 1mln tokens to an LLMs and waiting 6 min for the first token. 🀯

πŸ‘©β€πŸ’»πŸ§‘β€πŸ’» Check it out and contribute to our open-source project here: https://github.com/writer/writing-in-the-margins

🧠 More about chunked prefill: https://docs.vllm.ai/en/latest/models/performance.html#chunked-prefill
  • 2 replies
Β·
replied to melisa's post 3 months ago
view reply

We actually recently did an independent implementation of this paper in our open-source optimizing llm proxy optillm - https://github.com/codelion/optillm/blob/main/optillm/plugins/memory_plugin.py

We were able to use it as a basis for the memory plugin in optillm that gives LLMs short term memory. It helps improve accuracy on long context retrieval and even enables LLMs to have unbounded context if needed.

We were able to match SOTA on a recent benchmark from Google Frames benchmark (https://huggingface.co/datasets/google/frames-benchmark) with only gpt-4o-mini v/s Gemini 1.5 Flash which has a context length that is 10x more.

Screenshot 2024-10-04 at 8.25.41β€―PM.png
reacted to their post with πŸ‘€βž•πŸ€—β€οΈπŸ”₯πŸš€ 4 months ago
view post
Post
1944
We recently worked with OpenAI to fine-tune gpt-4o and built the SOTA model for the patched-codes/static-analysis-eval benchmark. All the code and data patched-codes/synth-vuln-fixes on how we did it is available on their GitHub - https://github.com/openai/build-hours/tree/main/5-4o_fine_tuning.

Here are some tips based on our experience:

β†’ Establish baseline with "conditioning" / prompting

β†’ Task-specific datasets are ideal for PEFT; hard to beat gpt-4o on "broad" tasks

β†’ Add your best system prompt to each example

β†’ Ensure training data distribution is similar to inference data

β†’ Shorten instructions with concise prompts; may require more examples.

β†’ Define clear evaluation metrics (seriously, please eval!)

You can see more details on the benchmark and process here - https://www.patched.codes/blog/the-static-analysis-evaluation-benchmark-measuring-llm-performance-in-fixing-software-vulnerabilities
posted an update 4 months ago
view post
Post
1944
We recently worked with OpenAI to fine-tune gpt-4o and built the SOTA model for the patched-codes/static-analysis-eval benchmark. All the code and data patched-codes/synth-vuln-fixes on how we did it is available on their GitHub - https://github.com/openai/build-hours/tree/main/5-4o_fine_tuning.

Here are some tips based on our experience:

β†’ Establish baseline with "conditioning" / prompting

β†’ Task-specific datasets are ideal for PEFT; hard to beat gpt-4o on "broad" tasks

β†’ Add your best system prompt to each example

β†’ Ensure training data distribution is similar to inference data

β†’ Shorten instructions with concise prompts; may require more examples.

β†’ Define clear evaluation metrics (seriously, please eval!)

You can see more details on the benchmark and process here - https://www.patched.codes/blog/the-static-analysis-evaluation-benchmark-measuring-llm-performance-in-fixing-software-vulnerabilities
reacted to their post with πŸ€—πŸš€πŸ”₯ 6 months ago
view post
Post
2504
A new paper titled "STALL+: Boosting LLM-based Repository-level Code Completion with Static Analysis" shows the benefits of integrating static analysis with LLMs. (https://arxiv.org/abs/2406.10018)

Authors evaluate 4 key questions:

- How does each static analysis integration strategy perform in LLM-based repository-level code completion?
> They found that integrating static analysis in the prompting phase (especially with file-level dependencies) can achieve the substantially larger improvements than other phases.

- How do different combinations of integration strategies affect LLM-based repository-level code completion?
> Languages that are easier to analyze like Java show more improvements compared to dynamic languages like Python.

- How do static analysis integration strategies perform when compared or combined with RAG in LLM-based repository-level code completion?
> Static analysis and RAG are complementary and boost the overall accuracy.

- What are the online costs of different integration strategies in LLM-based repository-level code completion?
> Combining prompting-phase static analysis and RAG is the best option for cost-effectiveness.

In my @owasp App Sec keynote last year, I had described how one can do static analysis augmented generation (SaAG) to boost the accuracy of LLM based patches for vulnerability remediation. (you can see the talk here - https://www.youtube.com/watch?v=Cw4-ZnUNVLs)
posted an update 6 months ago
view post
Post
2504
A new paper titled "STALL+: Boosting LLM-based Repository-level Code Completion with Static Analysis" shows the benefits of integrating static analysis with LLMs. (https://arxiv.org/abs/2406.10018)

Authors evaluate 4 key questions:

- How does each static analysis integration strategy perform in LLM-based repository-level code completion?
> They found that integrating static analysis in the prompting phase (especially with file-level dependencies) can achieve the substantially larger improvements than other phases.

- How do different combinations of integration strategies affect LLM-based repository-level code completion?
> Languages that are easier to analyze like Java show more improvements compared to dynamic languages like Python.

- How do static analysis integration strategies perform when compared or combined with RAG in LLM-based repository-level code completion?
> Static analysis and RAG are complementary and boost the overall accuracy.

- What are the online costs of different integration strategies in LLM-based repository-level code completion?
> Combining prompting-phase static analysis and RAG is the best option for cost-effectiveness.

In my @owasp App Sec keynote last year, I had described how one can do static analysis augmented generation (SaAG) to boost the accuracy of LLM based patches for vulnerability remediation. (you can see the talk here - https://www.youtube.com/watch?v=Cw4-ZnUNVLs)
reacted to their post with πŸ‘πŸ”₯πŸ€―πŸ‘€ 6 months ago
view post
Post
2231
LLM-Assisted Patching of Polyfill Supply Chain Attack

A recent supply chain attack on polyfill.io affected over 100,000 websites (see https://www.patched.codes/blog/patching-the-polyfill-supply-chain-attack). To address this issue, we show how developers can leverage Large Language Models (LLMs) for efficient vulnerability patching:

1. Automated Detection: Using Semgrep rules (see https://semgrep.dev/playground/r/KxUvD7w/asankhaya_personal_org.polyfill-compromise-copy) to identify vulnerable code.

2. LLM-Powered Patching: Utilizing Patchwork (https://github.com/patched-codes/patchwork), an open-source solution that employs LLMs to automatically fix vulnerabilities.

3. Custom Workflows: The "Fixpolyfill" patchflow (https://github.com/patched-codes/patchwork-configs/tree/main/patchflows/Fixpolyfill) , tailored for this specific attack, can be easily run across multiple repositories.

4. Scalable Solutions: Options to scan and patch entire GitHub/GitLab organizations, with automated pull request generation.

5. Rapid Response: LLM-assisted patching enables swift action to minimize damage from supply chain attacks.

This approach demonstrates how LLMs can be effectively used to quickly respond to and remediate widespread security vulnerabilities in code.
posted an update 6 months ago
view post
Post
2231
LLM-Assisted Patching of Polyfill Supply Chain Attack

A recent supply chain attack on polyfill.io affected over 100,000 websites (see https://www.patched.codes/blog/patching-the-polyfill-supply-chain-attack). To address this issue, we show how developers can leverage Large Language Models (LLMs) for efficient vulnerability patching:

1. Automated Detection: Using Semgrep rules (see https://semgrep.dev/playground/r/KxUvD7w/asankhaya_personal_org.polyfill-compromise-copy) to identify vulnerable code.

2. LLM-Powered Patching: Utilizing Patchwork (https://github.com/patched-codes/patchwork), an open-source solution that employs LLMs to automatically fix vulnerabilities.

3. Custom Workflows: The "Fixpolyfill" patchflow (https://github.com/patched-codes/patchwork-configs/tree/main/patchflows/Fixpolyfill) , tailored for this specific attack, can be easily run across multiple repositories.

4. Scalable Solutions: Options to scan and patch entire GitHub/GitLab organizations, with automated pull request generation.

5. Rapid Response: LLM-assisted patching enables swift action to minimize damage from supply chain attacks.

This approach demonstrates how LLMs can be effectively used to quickly respond to and remediate widespread security vulnerabilities in code.
reacted to their post with 🧠 6 months ago
view post
Post
7698
The new Claude Sonnet 3.5 model from Anthropic AI has been getting good reviews on since last night. It is quite good at coding related tasks. We tried it on the Static Analysis Eval benchmark ( patched-codes/static-analysis-eval) which measures the ability of a LLM to fix vulnerabilities. The model scores 59.21% which is good but not better than other frontier models (like GPT-4, Gemini-1.5 and LLama-3).
Β·