7 papers
SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasks
Seongho Son, Sangwoong Yoon, Jiahua Tang +3
Large language models (LLMs) embedded in multi-turn agentic harnesses are reshaping software engineering (SWE), but routing every task to a frontier model is wasteful when many iss…
LLM-WikiRace Benchmark: How Far Can LLMs Plan over Real-World Knowledge Graphs?
Juliusz Ziomek, William Bankes, Lorenz Wolf +3
We introduce LLM-Wikirace, a benchmark for evaluating planning, reasoning, and world knowledge in large language models (LLMs). In LLM-Wikirace, models must efficiently navigate Wi…
Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier
Lorenz Wolf, Connor Watts, Roger Creus Castanyer +4
The limiting resource for training agents via reinforcement learning (RL) is increasingly frontier task supply: valid, solvable tasks just difficult enough to train the current mod…
unix-ctf: Procedural Environments for Unix-Competence Reinforcement Learning
Geoffrey Bradway, Roger Creus Castanyer, Lorenz Wolf +3
Unix competence is the ability to use shell and operating-system primitives as first-class tools, not merely to write programs through a terminal. Current terminal benchmarks tend…
PopuLoRA: Co-Evolving LLM Populations for Reasoning Self-Play
Roger Creus Castanyer, Geoffrey Bradway, Lorenz Wolf +3
We introduce PopuLoRA, a population-based asymmetric self-play framework for reinforcement learning with verifiable rewards (RLVR) post-training of LLMs. Teachers and students are…
Reward Model Overoptimisation in Iterated RLHF
Lorenz Wolf, Robert Kirk, Mirco Musolesi
Reinforcement learning from human feedback (RLHF) is a widely used method for aligning large language models with human preferences. However, RLHF often suffers from reward model o…