The article presents GAG, a third‑generation framework that injects proprietary domain knowledge into frozen large language models using a single token, eliminating retrieval, avoiding base model updates, and maintaining constant inference budget while delivering strong performance on private QA and public benchmarks.
This article explains model editing—injecting or updating knowledge in AI models—distinguishes it from post‑training, outlines reliability, generalization and locality metrics, and surveys both parameter‑free (e.g., IKE) and parameter‑based methods such as ROME, hypernetworks, and MEND, highlighting practical challenges.
This article provides a comprehensive guide to customizing pre‑trained large language models through fine‑tuning techniques—including parameter‑efficient methods, data preparation, knowledge injection, and robust evaluation—offering practical steps, best practices, and domain‑specific considerations for achieving superior task performance.
This report details how large language models (LLMs) were integrated into Tencent's search advertising pipeline—from early extraction‑distillation experiments in 2023 to a 2024 end‑to‑end generative recall architecture—showing significant improvements in relevance, diversity, and revenue through knowledge injection, supervised fine‑tuning, constrained beam‑search decoding, and high‑performance inference services.
This article reviews the hallucination problem in large language models, analyzes its data, training, and inference sources, and presents Task‑aware Decoding (TaD) and Retrieval‑Augmented Generation (RAG) as effective, plug‑and‑play solutions demonstrated through extensive experiments.
This article reviews a recent study that bridges the knowledge gap between large language models and recommendation systems by generating natural‑language auxiliary tasks, fine‑tuning the models, and achieving notable performance gains on Amazon domain benchmarks.
The article describes Sunshine Insurance Group's Zhèngyán Large Model Open Platform, detailing its three‑layer architecture, AutoTrain tool, self‑developed LLM, smart routing, plugin marketplace, intelligent review, and how these capabilities empower insurance marketing, sales, service, and management through AI‑driven solutions.
KECP, a Knowledge‑Enhanced Contrastive Prompt‑tuning model, achieves strong few‑shot extractive question answering by converting questions to masked statements, injecting external knowledge via gated fusion, and leveraging contrastive learning alongside masked language modeling, as demonstrated on EMNLP‑2022 benchmarks.
This article reviews Alibaba DAMO‑ConvAI’s work on large‑scale conversational AI, comparing pretrained language and dialogue models, introducing the SPACE semi‑supervised knowledge‑injection framework and the Proton probing‑tuning method for extracting and applying model knowledge to downstream tasks.
