What You Say = What You Want? Teaching Humans to Articulate Requirements for LLMs
Paper
•
2409.08775
•
Published
Memory
Prompt is text-based memory. System II prompting is updating memory. Parametric memory is long-term, while prompt-based are short-tem.
-
-
OmniQuery: Contextually Augmenting Captured Multimodal Memory to Enable Personal Question Answering
Paper • 2409.08250 • Published • 1 -
Synthetic continued pretraining
Paper • 2409.07431 • Published • 1 -
WonderWorld: Interactive 3D Scene Generation from a Single Image
Paper • 2406.09394 • Published • 1
xRAG: Extreme Context Compression for Retrieval-augmented Generation with One Token
Paper • 2405.13792 • Published • 1Note Projecting document chunk embedding vector directly into hidden space for xRAG ! Explicit memory is expensive and dumb for RAG, mid-term memory relies on a 'projector', long-term memory updates on the langauge decoding part of the model. I guess that could be the next step here.
-
Evolution of Heuristics: Towards Efficient Automatic Algorithm Design Using Large Language Model
Paper • 2401.02051 • Published • 1
Yo'LLaVA: Your Personalized Language and Vision Assistant
Paper • 2406.09400 • Published • 1Note Addition of new concept into VLM via soft-prompt tuning. Extra id token in vocabulary plus k visual feature embeddings enables customizing VLM towards personalized knowledge.
LLMs + Persona-Plug = Personalized LLMs
Paper • 2409.11901 • Published • 15Note Same thing. Personalization with soft-prompt user embedding, this one is on text-modality, less exciting than Yo'LLaVA in some sense.