Topics / #in-context-learning

In-context-learning projects on GitHub

Capability of large language models to pick up tasks from examples shown in the prompt.

As of , the most-starred In-context-learning projects on GitHub are Foundations-of-LLMs (18.3K stars), Awesome-Multimodal-Large-Language-Models (18.1K stars) and LLMSurvey (12.2K stars). GitHub.Locker tracks 5 In-context-learning repositories with 49.3K stars combined. The fastest-growing on October 17 was Foundations-of-LLMs, with +5 stars in a day.

5Projects
49.3KStars in total
+11Stars in a day
Browse in the catalog →github.com/topics/in-context-learning ↗

Hot right now

Most stars gained in a day · Oct 17

+5

Foundations-of-LLMs

ZJU-LLMs/Foundations-of-LLMs · ★ 18.3K

An open-source textbook on the foundations of large language models, written by a team at Zhejiang University and freely available as a PDF.

+4

RoboICL

Mosi-AI/RoboICL · ★ 271

RoboICL is a research framework that lets a frozen, general-purpose vision-language model (GPT-6 Astra) control a robot in simulation purely through in-context learning, without training a dedicated robot policy.

+2

LLMSurvey

RUCAIBox/LLMSurvey · ★ 12.2K

It is the official GitHub page for the survey paper 'A Survey of Large Language Models' — a curated collection of papers and resources on how LLMs are built, adapted, and used.

The all-time top

5 projects · by stars

  1. #1 · ZJU-LLMs/Foundations-of-LLMs

    Foundations-of-LLMs

    An open-source textbook on the foundations of large language models, written by a team at Zhejiang University and freely available as a PDF.

    18.3K
  2. #2 · BradyFU/Awesome-Multimodal-Large-Language-Models

    Awesome-Multimodal-Large-Language-Models

    A curated, continuously updated list of papers, models, datasets, and benchmarks tracking the latest advances in Multimodal Large Language Models (MLLMs) — models like GPT-4o, Gemini, and VITA that understand images, video, and audio alongside text.

    18.1K
  3. #3 · RUCAIBox/LLMSurvey

    LLMSurvey

    It is the official GitHub page for the survey paper 'A Survey of Large Language Models' — a curated collection of papers and resources on how LLMs are built, adapted, and used.

    12.2K
  1. 4

    Zero-WAM

    434robbyant-research/Zero-WAM · Python

    Zero-WAM is a research method for zero-shot cross-task robotic manipulation: a causal video-action policy that learns in-context world-action modeling, predicting future robot observations and executable actions from either language instructions or human demonstration videos.

  2. 5

    RoboICL

    271Mosi-AI/RoboICL · Python

    RoboICL is a research framework that lets a frozen, general-purpose vision-language model (GPT-6 Astra) control a robot in simulation purely through in-context learning, without training a dedicated robot policy.

Often tagged together