Search papers, labs, and topics across Lattice.
IBM Research
5
0
5
9
Sequential testing can cut model evaluation costs by 80% while ensuring reliable results, challenging the status quo of fixed-size benchmarks.
Agents that excel on traditional benchmarks may crumble under the pressure of newly synthesized tasks, revealing the limitations of current evaluation methods.
Stop re-running full benchmarks: Calibrate new LLM datasets against existing suites with just 100 "anchor" questions and still get highly accurate performance predictions.
AI agents are far better at automating data engineering tasks than previously thought, but flawed benchmarks are obscuring their true potential.
General-purpose agents can match the performance of specialized agents across diverse environments without any environment-specific tuning, challenging the need for task-specific engineering.
