DARLING, a paper fixing repetitive LLM outputs after post-training, accepted at NeurIPS
DanielKhashabi · x · 2026-09-25
The DARLING paper from Daniel Khashabi's group has been accepted to NeurIPS. N8Programs calls it one of his favorite papers and the reason he got excited about joining JHU CLSP.
The problem it tackles: post-training tends to make language models repetitive. DARLING rewards responses that are both high quality and meaningfully different from one another, using a diversity-aware reward to counter mode collapse. Worth a look for anyone tracking post-training diversity.
More from Research
- Tuva Labs takes 1st on ARC-AGI-2, 2nd on ARC-AGI-3 in ARC Prize 2026 — GregKamradt · 2026-11-11
- Sakana AI × Pathway Post-Transformers lecture on AI and the brain now online — kaixhin · 2026-11-11
- Defending METR: public data can't second-guess its AI speedup findings — peterwildeford · 2026-11-11
- Adversarial RL envs for backdoor insertion vs. detection: where does it converge? — DimitrisPapail · 2026-11-11
- DiffusionEngine: an open-source modular PyTorch DDPM framework seeking contributors — Euphoric-Incident-93 · 2026-11-11
- UniPath (NeurIPS 2026): adaptively coordinating understanding and generation in one model — jindong_wang92 · 2026-11-11