The algorithm minimizes an advantage-weighted prediction loss, which maximizes reward, regularized by the rollout policy, which convexifies the objective and makes its optimization stable, and compares it to both forward- and reverse-process RL algorithms.
Branislav Kveton, Anup B. Rao, Subhojyoti Mukherjee et al.· 0 citations
Inspired by distributed computing frameworks like Apache Spark, DISCO partitions long context across a fleet of Worker LLMs dedicated exclusively to parallel, localized grounding, establishing a highly efficient paradigm for robust long-context inference.
Meta-Harness is presented, a system that casts harness design as search over three per-domain objectives (accuracy, behavioural safety, and token cost) solved by an agentic proposer (Claude Code) with full filesystem access to prior harness source, execution traces, and scoring artifacts.
A token-level probability reweighting framework that, given access to logits and a small amount of task-specific data, can effectively steer black-box LLMs toward application-specific content generation and advocates for broader access to token logits in closed-source models.
G. Hiranandani, Hao-Lun Wu, Subhojyoti Mukherjee et al.· International Conference on...· 7 citations· ⚡1
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.