Mihaiii

semantic-autocomplete

A blazing-fast semantic search React component. Match by meaning, not just by letters. Search as you type without waiting (no debounce needed). Rank by cosine similarity.

llm_steer

Steer LLM outputs towards a certain topic/subject and enhance response capabilities using activation engineering by adding steering vectors

detect-pretrain-code-contamination

laser

The Truth Is In There: Improving Reasoning in Language Models with Layer-Selective Rank Reduction

CommonGen-Eval

Evaluating LLMs with CommonGen-Lite

fastRAG

Efficient Retrieval Augmentation and Generation Framework

guidance

A guidance language for controlling large language models.

Homework

ic-wall

Crossover Ethereum/Internet Computer dApp demo and proof of concept. Sign in using Metamask, link eth address to IC identity FTW! ∞

juicy-draggable

Custom element to make items draggable and droppable

juicy-markdown-editor

GitHub Flavored Markdown (GFM) editor with filedrop and paste functionality

juicy-markdown-tabbededitor

GitHub Flavored Markdown (GFM) editor with filedrop, paste functionality and a preview window

juicy-popover

Custom Element that expands an area when clicked on a handle

KitchenSink

KitchenSink_

LearningClojure

llama.cpp

Port of Facebook's LLaMA model in C/C++

LLamaSharp

C#/.NET binding of llama.cpp, including LLaMa/GPT model inference and quantization, ASP.NET core integration and UI.

palindrom-redirect

Custom Element that redirects to a new URL when an attribute is changed. It's also being listened by Palindrom.

paper-search-input

An element providing a paper-input-like search, inspired by the search element on https://customelements.io

paper-tags

paper tags, input tags and paper-tags dropdown for selecting multiple items in a list

paper-tree

Browsable tree of nodes with expandable/collapsible capabilities and actions menu

RoAlpaca

Finetuning InstructLLaMA with portuguese data

RoverNavigator

Homework assignment made on 07.09.2021. Had 3h to complete.

sparsegpt-for-LLaMA

Code for the paper "SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot" with LLaMA implementation.

Tabbededitor

Intermediate work, nothing to see here, move along

TeamPulseRecentWorkItemsMailSender

tennis_MatchChartingProject

Raw, user-submitted point-by-point data for pro tennis matches

tennis_slam_pointbypoint

Point-by-point data for Grand Slams, 2011-15

translate