DataSense-Bench is a new benchmark that tests whether advanced AI models can choose and rank training data subsets that will improve a small language model’s performance. The benchmark lets agents inspect data, run analysis code, and perform forward passes, but not train the model or access evaluation tasks. After agents select subsets, the base model is fine‑tuned on each subset and evaluated, measuring both the performance of the best subset and the accuracy of the agents’ rankings. Experiments show limited gains over random selection and inconsistent ranking ability across tasks, with agents often using similar data signals but interpreting their value differently.
arXiv:2609.40097v1 Announce Type: new Abstract: Existing auto-research benchmarks often entangle multiple sources of improvement, including training frameworks, hyperparameters, compute budgets, and...
MLExpert is a 35B Mixture-of-Experts agentic model designed to autonomously perform long‑horizon machine learning engineering (MLE) tasks. It uses a post‑training method that combines hierarchical agentic supervised fine‑tuning with Windowed Interaction RL to learn from intermediate outcomes in complex experimentation workflows. On the MLE‑Bench Lite benchmark, MLExpert raises the average medal rate from 37.9% to 72.7%, matching the performance of larger frontier models and demonstrating generalizable optimization capabilities across related tasks.
EvalResearchBench (ERB) is a benchmark that tests whether AI agents can autonomously design evaluation tasks for recursive self‑improvement. In ERB, a researcher agent selects or synthesizes tasks, builds graders, and refines them through pilot tests before producing a final evaluator that runs within fixed time and API budgets. Experiments with nine researchers and thirteen candidate models show that the best automated evaluators achieve about 75% concordance with target benchmarks, below the 91% ceiling set by target disagreements, and that human‑designed task samples still outperform many automated approaches.
MLExpert is a 35B Mixture-of-Experts agentic model designed to autonomously conduct long‑horizon machine learning engineering (MLE) tasks. It uses a post‑training approach that combines hierarchical supervised fine‑tuning with Windowed Interaction RL to learn from intermediate outcomes in complex experimentation workflows. On the MLE‑Bench Lite benchmark, MLExpert raises the average medal rate from 37.9% to 72.7%, matching the performance of larger frontier models and demonstrating transferable optimization capabilities to other benchmarks such as NatureBench Lite.
The paper investigates whether large language models (LLMs) can anticipate the outcomes of machine‑learning experiments they generate. Using a dataset of 288 LLM‑proposed ideas across eight MLE‑bench tasks, the authors prompt five state‑of‑the‑art LLMs to predict each idea’s impact before training. Results show that while predictions outperform random chance, accuracy is limited—Claude Opus 5.5 explains only about half the variance (R² = 0.49), all models underestimate effect sizes, and performance varies by task, with coding harness choice having little influence.
The paper introduces SAEScientist-Bench, a benchmark that tests whether AI agents can autonomously conduct mechanistic interpretability research using Sparse Autoencoders (SAEs). Agents are tasked with designing contrastive probes and navigating a large feature dictionary in Gemma-2-9B-IT to identify optimal features for a target concept, with performance measured against expert-curated references on activation rank, concept selectivity, and causal steering. Results show that while frontier agents can discover features and outperform controls, they still lag behind expert baselines, especially in causal steering, highlighting both the potential and current limitations of closed-loop autonomous AI research.
arXiv:2607. 20465v1 Announce Type: new Abstract: The quality of training data fundamentally determines the capabilities of large language models (LLMs), yet no unified benchmark exists to measure how well LLMs, agents, and data-centric workflows actually prepare training data end to end.
OpenDiscoveryTrace is a public dataset of 558 complete AI scientific agent trajectories that records the reasoning process—thoughts, tool calls, observations, errors, revision triggers, and confidence—across 124 scientific tasks in drug discovery, materials science, genomics, and literature analysis. The dataset includes seven models (three frontier models and four open‑weight models) and 60 live‑retrieval variants, providing a balanced view of performance and error patterns. Pilot analysis shows that process traces reveal behavioral differences invisible to output‑only evaluation, such as differing error rates and types among frontier models.
Data World proposes a multi‑agent engine that generates synthetic data environments, tasks, and evaluation rubrics by first specifying the underlying world rather than curating real files. The system produces thousands of training tasks across diverse domains and enables independent audits of target recoverability, while human curation ensures realistic question selection. Benchmark results show that models trained on these synthesized tasks still face significant challenges, yet RL‑zero training on the synthetic data yields measurable performance gains over baseline models.
arXiv:2607. 06764v1 Announce Type: new Abstract: Recent progress on ARC-AGI-1 from disclosed architectures has come broadly from two regimes: heavy test-time compute over frontier models (evolutionary search, exhaustive sampling, extended chain-of-thought), or benchmark-specific training in which small models are fine-tuned on ARC data, often with task-specialized architectures.
arXiv:2608. 13940v1 Announce Type: new Abstract: AI research agents (AIRA) can now propose, implement, and evaluate their own machine learning experiments, but progress on frontier tasks is throttled by cost: a candidate solution can be written in minutes, whereas evaluating it can take hours to days of GPU time.
arXiv:2609.01244v1 Announce Type: new Abstract: Every supervised fine-tuning run forces the same chain of decisions, such as learning rate, batch size, LoRA or full fine-tuning, how many epochs, whic...