Tag

orcarouter/Qwen3.8-27B-Uncensored

Provides an abliterated (refusal-removed) build of Qwen3.8-27B for offline research and red‑teaming, keeping multimodal vision, an MTP speculative head, and a 262,144-token context. It has no built-in safety guardrails and is released under Apache‑2.0 for research use only.

Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL

Uses cooperative multi-agent RL where multiple decoupled models provide peer-derived pseudo-rewards to each other, enabling unsupervised improvements in reasoning; increases cohort diversity to reduce correlated errors and avoid training collapse, showing consistent gains across text and multimodal benchmarks.

Embodied-Navigator: Point, Think, Memorize, and Align for Efficient Navigation

Turns embodied navigation into 2D visual prompting where a vision-language model selects image pixels that are projected to 3D actions; adds selective chain-of-thought, compressed anchor-trajectory memory, and a two-level alignment objective to improve sample and runtime efficiency.

Agentic ESOpt: Fine-Tuning Long-Horizon LLM Agents with Minimal GPU Requirements

Fine-tunes long-horizon LLM agents with evolution strategies so full-model updates run at inference-level GPU memory. Emphasizes trajectory-level credit via black-box rewards, online prompt–parameter co-evolution, and a cosine decay for perturbation scale to balance exploration and adaptation; suited for limited-GPU settings.

Ornith-1.5-35B-A3B

A 35B mixture-of-experts LLM tuned for agentic coding and end-to-end self-improvement: it jointly generates tasks, scaffolds, and solution rollouts. Activates ~3B params/token, supports 256K context (extendable), and emits chain-of-thought plus OpenAI-style tool calls.

Ornith-1.5-35B-A3B-GGUF

GGUF build of Ornith-1.5's 35B mixture-of-experts model (A3B) for local inference — activates ~3B params per token, supports up to 262,144 tokens, emits separate reasoning traces and OpenAI-style tool calls, optimized for agentic coding and long-context use cases.

Qwen3.8-27B-DFlash2

A draft model that predicts whole blocks of tokens in parallel for speculative decoding of Qwen3.8-27B. Uses block-diffusion drafting with per-position candidate sets and a selector plus dynamic convolutions to keep end-of-block accuracy, increasing accepted tokens per verification and end-to-end throughput versus autoregressive decoding.

Ornith-1.5-9B-GGUF

A 9B dense reasoning LLM optimized for single‑GPU deployment and terminal-based coding agents, with long-context support (up to 262,144 tokens) and GGUF/quantized builds for edge/mobile. Strong on coding and agentic benchmarks.

Qwen3.8-27B-Escha-W2

Provides 2-bit quantized weights of Qwen3.8-27B (~10.15 GB) for local deployment, enabling the full 27B parameter model to run on a single 24 GB GPU with long-context support. Delivered as safetensors plus a companion SGLang runtime; measured to match FP8 reference on common benchmarks with small or no quality loss.

Graph Engineering in the Era of LLM Agents: From Individual Intelligence to System Intelligence

Proposes “Graph Engineering”: using explicit, dynamic graphs to represent tasks, agents, tools, and system state so LLM-based agent systems can coordinate, persist, and evolve. Surveys principles, methods, applications, and curates related resources.

Spark-X2.5-4B

A 4B-parameter on-device general-purpose LLM for chat, writing, translation, coding and agentic workflows with native 1,000,000-token context. Uses a hybrid attention design to enable long-context efficiency, pretrained on ~20T tokens, and compatible with vLLM, llama.cpp, Ollama and LM Studio.

IFM/Pretrain-Behaviors

Behavior-focused text corpus for LM pretraining, organized into seven Parquet-backed subsets (reasoning, planning, data-science, games, general, format-rewrites, other). Supports streaming, custom sampling, and large-scale dataset pipelines for research and model training.