AI Wiki — Knowledge Base

Connected a local model (Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M) to GIMP via MCP tools using llama.cpp - and here's the image result from my first prompt 'can you draw a picture of a flower in gimp?'. Needs work. Setup follows.

can you draw a picture of a flower in gimp? So I ran through this setup to install the MCP server suite from GITHUB, it's pretty easy to follow the install - and installs the libraries necessary to co

model-releasesr-localllama

16 Sept 2026

Search: “context”

Evaluating Communicative Success in Machine-Translated Conversation

G^2RA-NET: Graph-based Cross-Slice Relation Modeling with Attention Gating for Medical Image Segmentation

Generalization through Lexical Abstraction in Transformer Models: The Case of Functional Words

Instance Segmentation and Fine-grained Classification for Urban Buildings with Adaptive Region Dividing and Spatially-Supervised Contrastive Learning

Not All Layers Are Equal: Dynamic Layer Routing for Reliable CLIP OOD Detection

Qwen3.8-Flash-Next (95.5 GiB) on a 64GB Mac at ~27 tok/s, checkpoint + fork

Rethinking Multi-Agent Collaboration: When More Is Less

Self-State Attacks on Self-Hosted AI Agents: How Far Can OS Defenses Go?

SIMLIFE: Pattern Understanding for Long-Horizon Human-Agent Partnership

To Copy or Not to Copy: Controlling Speculative Decoding via Intrinsic Model Signals

Uni-LaDiR: Latent Diffusion Unifies Multimodal Reasoning

A Comprehensive Review of Generative Physical Artificial Intelligence

Behavior2Value: Benchmarking and Empowering LLMs for Consumer Value Measurement from E-commerce Behaviors

BENCHCOMPASS: From Scores to Signals for Training and Harness Decisions in Payment-Domain LLMs

Collaborative Memory for Multi-Agent VLM Systems

Consensus-based optimization for closed-box adversarial attacks and a connection to evolution strategies

Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches

Flag Game: A Toy Model for Mechanistic Swarm Interpretability

From a River in Gilead to the Inference Distributions of Large Language Models: Covert Dialect Bias and Linguistic Profiling at Scale

Information Set Emulation: Causal Certificates for AI Derived EHR Features

Knowledge-Graph Based Augmentation versus Retrieval Augmented Generation for Cultural-Related Question Answering

LocQE: Principled Domain Adaptation for Localisation Quality Estimation by Leveraging Post-Edits

Making Political Text Scaling Comparable: Infrastructure and Hyperparameter Sensitivity for 17 Algorithms

Newer Is Not Fairer: Gender Stereotyping in Text-to-Image AI Across Model Generations

Planning or Improvisation? Stress-Testing the Poetry Planning Site on Open Models and Open Cross-Layer Transcoders

Qwen 3.8 27b is a amazing model, for the first time I see a local model found its own away to open a browser and test

Re2A: Situated Conversational Recommendation via Rubric-based Preference Reasoning and Alignment

Rethinking How We Evaluate Methodological Progress in Health AI

Risk-Aware World Modeling with Flow-Guided Occupancy Evolution for Selective Trajectory Planning in Automated Driving

Symbolic Temporal Supervision of LLM Agents Using Contracts

T-SANDHI: Tone Sandhi-aware Adaptive Network with Decoupled Hybrid Injection for Low-resource Taiwanese Hokkien Speech Recognition

The Attention Within: Consensus Dynamics in Selective State Space Models

Time-Aligned Evolving Concept Graphs for Scientific Relation Forecasting

TRIPROBE: Probing Task Separability Beyond Classification for XAI

Two weeks. That's how long it took to go from GLM-5.3-Flash's first run on domestic accelerators to serving all of its production traffic, w…

v0.34.2-rc2: mlxrunner: Release freed KV buffers during speculative decode

Version- and Scope-Aware Question Answering over Normative Documents: A Deployed System and an End-to-End Evaluation at Production Scale

Affect-Prototype Guided Fusion for Open-Vocabulary Incomplete Multi-modal Emotion Recognition

AI for Games in the Foundation Model Era

Anatomy-Change-Aware Bidirectional Selective State-Space Memory for Clinically Deployed Thoracic Radiotherapy Auto-Contouring

AquiLLM: Evaluating Faithfulness in Open-Weight RAG-LLM Systems for Scientific Research

Claude Code showed that AI could do real work, not just answer questions. Developers hand Claude a feature, come back to shipped code. That'…

Connected a local model (Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M) to GIMP via MCP tools using llama.cpp - and here's the image result from my first prompt 'can you draw a picture of a flower in gimp?'. Needs work. Setup follows.

De-GAN - Dynamic Parameter Tuned GAN for 3D Medical Image Segmentation: A Step Towards Generalisation

Diagnosing the Fact-Grounding Gap in Multi-Hop Question Answering

Disrupted Companionship: A Risk Assessment Framework and Cross-Platform Quantitative Analysis of Psychosocial Responses to AI Companion Disruptions

DriveMCP: An Agentic AI framework for Advanced Driver Assistance System

Easy to Catch a Liar, Hard to Clear an Honest One: Language Models Diagnosing a Corrupted Reward Channel from a Verified Record