RL & Alignment

LoCo learns state-conditioned continue-or-conclude preferences from self-generated math traces by freezing overthinking prefixes and training DPO on shorter verified continuations, reducing reasoning length while preserving or improving accuracy.

Axion 1.0 is a fully open agent foundation model trained with two SFT stages plus RL, without dedicated mid-training, DPO, or OPD, and is competitive across a broad agentic benchmark suite.

NorthCUA is an integrated harness, experience-production flywheel, and staged-learning pipeline that trains Nano, Mini, and Pro computer-use agents and reports strong GUI grounding, desktop, web, and latency results.

R2Orch nests self-play orchestration with a learnability-based Proposer and randomized skill-withdrawal tests, improving a Director toward zero-skill deployment across twelve benchmarks.

SimpleLoRIF uses scalable per-example gradient embeddings to identify and remove benign-looking fine-tuning data whose gradients align with unsafe behavior, reducing jailbreak attack success while preserving capability.

WeFair trains a compact personalized short-video query rewriter by matching human-feedback fidelity to curriculum SFT, consumption-pair DPO, and strict-target GRPO, reducing intent drift and improving online engagement.

CaPtain is an open post-training pipeline (SFT on synthetic in-harness trajectories, then multi-turn RL with a privileged simulator critic) that turns a 4B open VLM into a code-as-policy robot agent operating over 22 low-level perception and motion primitives, reaching 73.8% on held-out LIBERO.

The paper splits training-inference mismatch into an approximately fixed numerical discrepancy and a self-induced policy lag, then proposes DLC, which centers the discrepancy, clips the lag ratio, and resynchronizes the sampler when the lag ratio's effective sample size falls below one half.

A-RSI uses training-log reward-variation and accuracy signals to decide when to replace a pool of programmatically generated multimodal tasks, and reports gains on an eleven-benchmark suite without external training examples or human answer labels.

AR-BED trains a reasoning LLM with information-gain rewards under an externally specified finite Bayesian experimental-design model, improving adaptive questioning and transferring to a clinical dialogue task.

LongZero lets an LLM improve long-context reasoning from zero external data by first writing structured question-evidence-answer specifications, expanding them into long multi-document contexts, adding self-written retrieval conversations, and training with GRPO.

V4.1 Flash-1017 shows strong safety improvements over DeepSeek V4 models in evaluation integrity, loss of control, and harmful manipulation, but retains weaknesses under sustained pressure, partial oversight, and assigned deceptive goals.