Compresses Qwen3.8-27B into a 12.3 GB sensitivity-aware mixed-precision quantized checkpoint for long-horizon agent workloads; preserves BF16 fidelity (+0.02% PPL, 93.2% token Top‑1 agreement), supports 262K context and vLLM serving, text-only and Apache‑2.0 licensed.
Provides a plug-and-play harness that makes existing agents omni-native by exposing hierarchical multimodal Skills, a standardized execution interface, dependency-aware orchestration, and a persistent Asset Registry. Represents multi-asset workflows as Declare Execution Graphs to schedule concurrent operations and enable cross-turn reuse across interchangeable execution backends.
Provides a high-fidelity mixed-precision (≈3-bit) GGUF quant of Qwen3.8-27B tailored for long-horizon agent workloads and cyber-focused red-teaming. Preserves reasoning, thinking mode, MTP speculative decoding and vision (via a separate mmproj); released as an uncensored/abliterated research build under Apache-2.0.
Investigates how rollout policy, token-level KL direction, and learning rate each affect LLM distillation across Llama3 and Qwen2.5 on reasoning tasks; finds KL direction and learning rate dominate outcomes while rollout policy has a modest effect.
Teaches small reasoning models to diagnose when additional internal thinking is insufficient and to selectively query stronger models; introduces FlyBy, a supervised + cost-aware RL framework that learns when/what to ask. Improves pass@ metrics on hard benchmarks while reducing serving cost.
Improves test-time scaling of looped transformers by adaptively assigning extra recurrent iterations to tokens that benefit most. TaH2 is a post-training method that jointly trains an iteration decider with the backbone using lookahead depth supervision, boosting the accuracy–compute slope and peak accuracy on challenging benchmarks.
Trains a single unified multimodal model with reinforcement learning to perform end-to-end self-reflection and iterative image repair — jointly learning the diagnostic (textual) reflection and the flow-based image revisions so credit propagates across rounds without an external verifier.
Defines and evaluates AREX-2, an LLM agent that iteratively self-improves at test time via reflection and long-horizon execution. Trained on long-horizon improvement trajectories from ML engineering and algorithmic programming (built on Qwen3.8-27B), it scales with more rounds and achieves strong benchmark scores.
Demonstrates that pretrained transformers typically use only ~1–3 lines of depth to follow reference chains, and that a task‑trained rank‑8 LoRA applied at one early layer (with all other weights frozen) can extend reference‑following to dozens or hundreds of lines while adding only a few ten‑thousand parameters.
Turns solved protein structures into FoldingCorpus and Fold2Reason — a post-training recipe that supervises an LLM with discrete structural Q&A plus continuous 3D geometry to improve spatial, graph and scientific reasoning; reports consistent gains across 10 benchmarks.
Processes text and images, with support for reasoning, coding, and tool use. Its trillion-parameter MoE activates about 49B parameters per token and supports a 1M-token context.
Provides an EXL3 3.0 bits-per-weight quantization of a weight-edited GLM-5.3 UNCENSORED FP8 model for self-hosted text generation and agent workflows. Key characteristics: 753B MoE architecture, 273 GiB on disk, converted with ExLlamaV3; tool-call parsing requires preserving string arguments.