inference-scaling

Diffusion-Sharpening: Fine-tuning Diffusion Models with Denoising Trajectory Sharpening

[NeurIPS 2024] Fast Best-of-N Decoding via Speculative Rejection

A Python library for inference-time scaling LLMs

Stable Latent Reasoning --- Enhancing Inference in Large Language Models through Iterative Latent Space Refinement

Implemented a recurrent-depth LLM (PyTorch) based on arXiv:2502.05171. Demonstrated that scaling inference compute increased arithmetic reasoning accuracy from 8% to 100% without additional parameters.

Do stacked LLM verifiers actually multiply reliability? An empirical-Bayes inversion that recovers verifier correlation (ρ_v) and the reliability ceiling from accept/reject logs — and falsifies the independence "Odds Law."

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

The Architectural Inversion of Machine Reasoning to Suit Its Suitors

Private AI on NeoClouds - Large Open Weight models need a lot of VERY large GPUs (aka massive infra)

ReasonForge — Interactive Test-Time Compute Laboratory. Live scaling curves, MCTS+PRM, Tree-of-Thoughts, Best-of-N verifier, Self-Refine reward-hacking — all on the Game-of-24 benchmark.

Hands-on demo mapping the accuracy-vs-token-cost Pareto front of LLM test-time scaling strategies: chain-of-thought, self-consistency, debate, and mixture-of-agents.

Investigating how inference-time compute influences reasoning performance in large language models through systematic scaling of self-consistency decoding.

LLM elevation protocol: make a daily-driver model (Sonnet) operate near frontier level by buying depth with inference compute — GEN×3, adversarial verification, external memory, and distilled-judgment cartridges. Claude Skills format, ready to install.

Deep Research capability with reasoning models, CoT prompting, and inference-time scaling