Benchmarks across 70+ tasks show cross-encoders deliver zero net gain, only latency wins
bo_wangbo · x · 2026-10-09
In a reply, bowangbo says his team benchmarked every "SOTA" cross-encoder from 200m to 8B parameters, point-wise to list-wise, across 70+ tasks and domains. After averaging gains and losses, all cross-encoders deliver zero net gain — the only benefit is reduced latency.
Related event: 70+ Task Benchmark Shows SOTA Cross-Encoders Yield Zero Average Gain(2 posts)→
More from Research
- EMNLP paper: diffusion LMs internally encode latent denoising timestep representations — gsarti_ · 2026-10-23
- KU Leuven and VIB open two Group Leader professorships in AI for Biology, start 2027 — jpillowtime · 2026-10-23
- Subspace Tuning curbs LLM catastrophic forgetting by freezing principal singular directions — scottniekum · 2026-10-23
- Perceptic launches PercepticOS, an AI workspace for 'scientific superintelligence' — nathanbenaich · 2026-10-23
- EPFL's Bosselut to talk long context as test-time learning at ELLIS seminar — andre_t_martins · 2026-10-23
- DRATS: Adaptive Task Sampling Boosts Worst-Task Performance in Multi-Task RL — Cohere · 2026-10-23