Search posts, papers, and topics
SoL-Pi is a new approach for scaling auto-research loops in coding agents, focusing on token efficiency for long reasoning trajectories. It achieves comparable performance to existing methods while significantly reducing token traffic by 44.7-49.0% and API costs by about one third on the 51-task EdgeBench evaluation.
The talk argues that prompt engineering is a dead‑end and proposes building large evaluation/optimization pipelines (pass^k testing, adversarial scenario generation, automated prompt optimization) to make LLM agents reliable. It describes a workflow: generate tests, run them with/without a new “skill”, feed results to a genetic optimizer that mutates prompts, validate on hold‑out tests, and itera…
A step‑by‑step tutorial showing how to use the open‑source scikit‑decide framework together with the OpenAP aircraft performance model to compute fuel‑optimal flight trajectories. The post details the author’s high‑end workstation, installs Python 3.12, scikit‑decide, OpenAP, and DuckDB with several extensions, then explores OpenAP’s aircraft data (e.g., A380‑800 specs and drag polar) and demonst…
The paper proposes Comparison-based Preference Optimization (ComPO), a zeroth‑order method that uses comparison oracles to align LLMs without a differentiable loss. Experiments on several LLM families show it improves win rates and mitigates likelihood displacement compared to direct alignment approaches.
The paper reframes fine‑tuning of instruction‑tuned LLMs as a direction‑selection problem under a fixed behavioral‑drift budget, showing that the update direction, not magnitude, determines trade‑offs between target performance and capability preservation. In QA‑only fine‑tuning of Qwen‑3 models, layer‑selective probing finds effective directions that boost scientific reasoning and multilingual t…
The authors cast transformer block removal as a constrained binary optimization problem equivalent to an Ising glass, using a Hessian‑derived energy as a proxy for downstream quality. Solving the resulting QUBO with classical or quantum‑inspired solvers yields up to 23 MMLU points improvement over prior block‑removal baselines at 50 % depth compression.
