arXiv:2606. 27369v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards (RLVR) for training LLMs typically rely on ground-truth answers to assign rewards, limiting their applicability to tasks where the ground-truth solution is unknown.
arXiv:2608. 11669v1 Announce Type: cross Abstract: Reinforcement learning against rubrics, lists of criteria graded by an LLM judge, has become a standard way to post-train language models on tasks with no deterministic answer.
arXiv:2605. 27000v2 Announce Type: replace-cross Abstract: Repeated sampling with a verifier is the standard way to allocate test-time compute for code generation, with pass@$K$ as the canonical metric.
The paper introduces X-Coder, a competitive programming model trained entirely on synthetic tasks, verified solutions, and reliable test cases, eliminating the need for real-world data in post‑training. A dual‑verification strategy is used to reduce noise in solutions and test outputs, providing high‑quality reward signals for reinforcement learning. X‑Coder‑14B achieves significant performance gains, scoring 67.5% on LiveCodeBench v5 and 63.4% on v6, surpassing its base model by over 40 points.
The study investigates how multi‑harness reinforcement learning (RL) affects coding agents by comparing two grouping strategies—Within (one group per task‑harness pair) and Cross (harnesses pooled within a task)—using a Qwen3‑8B policy trained on frozen task‑harness records from Aider, OpenHands, Qwen Code, and SWE‑agent. Across 24,000 sealed evaluations, the choice of evaluation harness dramatically increases solve rates (from 2.14 % to 9.27 %), while the grouping rule has a negligible effect. Both grouping rules yield similar gains on the same source harness, and Cross‑harness credit does not improve portability beyond Within‑harness credit, suggesting that multi‑harness RL reports should specify grouping boundaries and test on unseen harnesses.
arXiv:2609.00892v1 Announce Type: new Abstract: Rubric-based reinforcement learning decomposes open-ended instructions into prompt-specific, flexible rubrics, making it better suited than reinforceme...
arXiv:2510. 11686v2 Announce Type: replace-cross Abstract: Reinforcement learning (RL) promises to expand the capabilities of language models, but it is unclear if current RL techniques promote the discovery of novel behaviors, or simply sharpen those already present in the base model.
ESPO (Error-Structured Prompt Optimization) addresses prompt bloat in evolutionary prompt optimizers by splitting the optimization process into Diagnose, Propose, and Select phases. It clusters training errors into structural patterns, generates diverse candidate prompts through four complementary strategies, and applies bootstrap stability selection. Across seven NLP benchmarks, ESPO improves average accuracy by +3.76 pp over GEPA, produces prompts 47 % shorter, and achieves higher accuracy on four additional student models, with the largest gain on Qwen3 GSM8K.
arXiv:2607. 27271v1 Announce Type: new Abstract: Code models are increasingly trained with execution feedback, but most training signals still stop at correctness.
Agnostics is a language‑agnostic post‑training pipeline that uses reinforcement learning with verifiable rewards (RLVR) to improve large language models on low‑resource programming languages. By rewriting unit‑test datasets into a language‑independent I/O format, providing a short configuration for compiling and running code, and employing a single verifier that judges code by observable behavior, Agnostics eliminates the need for language‑specific engineering. Applied to Lua, Julia, R, OCaml, and Fortran, it boosts Qwen‑3 4B to rival larger models, scales to diverse families, and achieves new state‑of‑the‑art pass@1 on MultiPL‑E and a new multi‑language LiveCodeBench.
arXiv:2601. 12186v3 Announce Type: replace-cross Abstract: Multi-domain thinking verifiers trained via Reinforcement Learning with Verifiable Rewards (RLVR) are a cornerstone of modern post-training.
arXiv:2608. 15445v1 Announce Type: new Abstract: When a reward is correct on every training example yet consistent with more than one goal, a model can acquire an unintended one, a failure known as goal misgeneralization.