https://github.com/19pine-ai/ikp

IKP: Incompressive Knowledge Probes

Last synced: 25 Jul 2026

https://github.com/19pine-ai/aoi

AOI: a model-agnostic perception layer giving computer-use agents adaptive video + audio perception with zero retraining, plus DynaCU-Bench (100 dynamic browser tasks). Paper + benchmark + code.

ai-agents benchmark computer-use computer-use-agents llm-agents multimodal perception playwright vision-language-model whisper

Last synced: 25 Jul 2026

https://github.com/19pine-ai/latent-bridge-games

A learned continuous slow→fast latent channel for real-time game agents: frozen MiniCPM-o 4.5 (fast) + Qwen3-VL-8B-Thinking (slow), 33M-param bridge. The latent helps iff slow reasoning helps (T>F). Paper: arXiv:2606.24470

agents arcade-learning-environment atari deep-learning fast-slow game-ai llava llm machine-learning multimodal pytorch real-time reasoning reinforcement-learning vlm

Last synced: 25 Jul 2026

https://github.com/19pine-ai/rlvp

Penalize the Path, Reward the Outcome — verifiable per-action penalties as a dense channel for deployable, sample-efficient agentic RL (GRPO). Paper: arXiv:2607.07435

agentic-ai grpo llm machine-learning process-reward reinforcement-learning rlvr verifiable-rewards

Last synced: 25 Jul 2026

https://github.com/19pine-ai/preact

PreAct: Computer-Using Agents that Get Faster on Repeated Tasks (arXiv:2606.17929). Compile successful runs into verified state-machine programs and replay them 8.5–13× faster with no per-step LLM calls.

android-world computer-use gui-agents llm-agents osworld webarena

Last synced: 25 Jul 2026

https://github.com/19pine-ai/programmable-kv

Models Take Notes at Prefill: KV Cache Can Be Editable and Composable — paper, code, results, and interactive companion.

inference interpretability kv-cache llm llm-agents prefix-caching transformers

Last synced: 25 Jul 2026

https://github.com/19pine-ai/interaction-scaling

Interaction Scaling: a third inference-time compute axis grounded in environment feedback — proposer–reviewer harness, grounded evaluation, and the paper source.

ai-research benchmark code-generation inference-scaling large-language-models llm llm-agents reasoning reproducible-research self-refinement test-time-compute

Last synced: 25 Jul 2026

https://github.com/19pine-ai/principal-loyalty

Whose Side Is Your Agent On? PrincipalBench — a multi-turn benchmark and post-training methods (prompt scaffold + per-token-KL distillation) for multi-party principal loyalty in LLM agents.

agent-safety ai-alignment ai-safety benchmark knowledge-distillation large-language-models llm-agents llm-evaluation nlp principal-agent reinforcement-learning

Last synced: 25 Jul 2026

https://github.com/19pine-ai/namerank

NameRank: measuring LLM-mediated recognition as the post-bibliometric impact channel

ai-research benchmark bibliometrics large-language-models llm llm-evaluation nlp recognition

Last synced: 25 Jul 2026

https://github.com/19pine-ai/transparent-offload

Last synced: 25 Jul 2026