Removes the subspace of frequent, uninformative tokens that LLMs inject into text embeddings via the model's unembedding matrix. EmbedFilter is a lightweight linear transform that refines LLM-derived embeddings to improve zero‑shot semantic retrieval, enable dimensionality reduction, and speed up indexing; code on GitHub.
Analyzes the parameter-space geometry of on-policy distillation (OPD) for LLM training, showing OPD updates affect fewer weights, avoid principal directions, and rapidly lock into a low-dimensional update subspace. Compares OPD with supervised fine-tuning (SFT) and reinforcement learning with verifiable rewards (RLVR) and studies implications for optimization and objective mixing.
Models visual preference as distributions over rubric scores and introduces Z-Reward, a teacher–student framework that decouples reasoning-heavy judgment (teacher trained with GDSO) from efficient deployment (student via RISD). Demonstrates higher human-preference accuracy and works as a differentiable reward for text-to-image optimization.
AI Agent Papers2026
Lets a single LLM simultaneously act as agent and environment to bootstrap co-evolutional training — using state-prediction process rewards (World-In-Agent) and failure-mode retrieval (Agent-In-World) to reshape training data; reports ~4% average benchmark gain.
AI Model2026
Generates text from interleaved text, image, and short-video inputs using discrete diffusion and block‑autoregressive multi‑canvas sampling; built on a sparse MoE (8/128) Gemma 4 backbone and optimized for low‑latency inference and very long contexts (up to 256K tokens).
AI Model2026
Fine-tuned Hugging Face image-generation model that biases Ideogram-style prompts toward photorealistic outputs. Emphasizes natural lighting and realistic materials to reduce prompt tweaking; license not specified.
Proposes a router redesign for Mixture-of-Experts (MoE) that aligns each router row with its expert's principal singular direction using Manifold Power Iteration (MPI), improving token–expert affinity. MPI applies a 'power‑then‑retract' step to push router rows toward principal singular vectors while enforcing norm constraints; the paper gives convergence theory and pretraining results on 1B–11B MoE models.
AI Video Papers2026
Controllable long-horizon text/image-to-video generation that supports camera navigation, revisits, and promptable events across photorealistic and stylized domains. Introduces camera-aware positional encoding (E-PRoPE), memory-conditioned scene persistence, causal-forcing distillation, and RL alignment to retain camera control and reduce drift.
Language-conditioned robot policy that reuses a pretrained geometric foundation model and inserts a causal future predictor at an intermediate layer so the same backbone produces future 3D-aware features and action outputs, enabling geometry-aware temporal prediction with minimal architectural change.
AI Model2026
Provides a large language model optimized for long-horizon agentic tasks and end-to-end coding workflows — with a stable 1,000,000-token context, IndexShare sparse-attention and multi-level thinking-effort modes. MIT-licensed and designed for deployments that need sustained long-context reasoning and coding.
AI Model2026
Provides FP8-quantized weights of GLM-5.2 — a 744B long-context LLM tuned for sustained 1M-token engineering, coding and agentic workflows; compatible with vLLM, Transformers, SGLang and Ascend NPU deployments.
AI Image2026
Generates images from natural-language prompts as an 8-step distilled checkpoint of Krea 2, optimized for fast iterative text-to-image workflows with style references and 1K–2K resolution outputs.