Tag
Explore by tags
Drop-in Jinja chat templates for Qwen 3.5/3.6 that fix rendering errors, token waste, and tool-calling failures across runtimes (LM Studio, llama.cpp, vLLM, MLX). Adds a think-on/think-off toggle, auto-closes broken thinking tags, robust tool-argument handling, and a graceful fallback for missing user queries.
Supervised fine-tuning dataset of 7,716 reasoning-focused Q&A examples distilled from the DeepSeek‑V4‑Flash teacher; provided as a cleaned JSONL train split for distillation and SFT experiments.
Generates page-scale UI designs and audits for Claude Code, Cursor, and Codex using a 57-gate “anti-AI-slop” rule set — produces distinct, non-template HTML+CSS outputs and supports audit, redesign, and study verbs with a built-in pre-emit self-critique.
Early-preview (≈1.2k rows) dataset of agentic coding prompts and unedited model responses generated by DeepSeek‑V4‑Pro, covering real-world programming tasks across many languages. Intended for research, filtering, and model evaluation rather than production training without review.
Instruction‑tuning dataset of 8,706 Claude Opus 4.6/4.7–generated examples where each assistant turn begins with a synthetic <think> block to emulate chain‑of‑thought. Provided as four splits (full/instruct/roleplay/code), ~17M tokens total, Apache‑2.0, not manually reviewed.
A prompt-only mixture of ~478k prompts designed to support antidoom-style generation and preference-data pipelines for reducing model repetition (doom loops). Prompts are stripped of answers and labels and sourced from many public datasets so it’s usable for FTPO/adapter generation but not for supervised QA evaluation.
Provides 19,331 multi-turn ChatML Hermes reasoning traces produced by DeepSeek V4 Pro for LoRA fine-tuning of agent-style models; includes VRAM-tiered variants, train/valid/test splits, and dense tool-calling annotations in Parquet format.
Provides 19,331 multi-turn ChatML Hermes reasoning traces for LoRA fine-tuning of local models to behave as Hermes agents. Includes train/valid/test splits, VRAM-tiered variants (nano→spark), ~138K tool-call annotations, and Parquet format under Apache-2.0.
Agentic coding evaluation dataset containing real-world, multi-step developer tasks and raw model responses across 20+ programming languages. Emphasizes challenging, persona-driven prompts for benchmarking and fine-tuning; users should filter and audit outputs before training.
Trains reusable natural-language 'skills' for frozen LLM agents by optimizing the skill document in text-space — using trajectory-driven edits, validation-gated updates, and deployable best_skill.md artifacts. Multi-backend, zero inference-time cost at deployment, designed for iterative, validation-led skill improvement.
Generates and edits high-resolution images (up to 2048×2048) from text and reference images, plus subject-driven personalization. Implements a pixel-level unified transformer that encodes raw pixels and text in one token space and includes a reasoning-driven prompt agent for layout and text rendering.
Provides a locally runnable 26.9B Qwen3.6 checkpoint that surgically reduces refusal behavior in weight space while preserving capability; ships bfloat16 safetensors and a GGUF quant ladder for local runtimes and red-team evaluation.