arXiv:2606. 04161v1 Announce Type: new Abstract: Different predictors often excel on different inputs, so picking the best one per instance promises higher accuracy than committing to a single model.
arXiv:2606. 29280v1 Announce Type: cross Abstract: We identify intervention bias as a previously unquantified failure mode of zero-shot large-language-model (LLM) educational advisory agents: without task-specific training, they recommend action when a hindsight-optimal oracle policy mandates inaction.
arXiv:2609.01244v1 Announce Type: new Abstract: Every supervised fine-tuning run forces the same chain of decisions, such as learning rate, batch size, LoRA or full fine-tuning, how many epochs, whic...
arXiv:2606. 30852v1 Announce Type: new Abstract: Reasoning models spend different amounts of useful computation across instances, but it remains unclear when a learned stopping rule improves over simple confidence or convergence thresholds.
Every supervised fine-tuning run forces the same chain of decisions, such as learning rate, batch size, LoRA or full fine-tuning, how many epochs, which optimiser, and what data to feed the model. Eac...
arXiv:2607. 17531v1 Announce Type: cross Abstract: Test-time collaboration, including self-consistency, best-of-N selection, critic models, and verifier pipelines, is often credited with broadly improving LLM reasoning, yet its gains are uneven and sometimes negative.
The paper proposes a method for selectively querying language‑model advice in reinforcement learning by predicting the value of potential responses and only querying when the expected benefit outweighs the cost. It introduces a certified, response‑contingent metareasoning framework that guarantees near‑optimal advice usage under certain assumptions, and demonstrates that a calibrated controller with Qwen2.5 advisors can improve task performance while drastically reducing the number of advice calls on the BabyAI benchmark.
arXiv:2608. 02636v1 Announce Type: cross Abstract: Self-evolving skill systems promise to improve agents by turning execution feedback into persistent skill updates without changing the underlying model.
arXiv:2608. 15565v1 Announce Type: new Abstract: Experience-learning agents for optimization modeling improve by storing verified skills, but existing learners admit knowledge by checking against known answers, which real ticket streams do not provide.
arXiv:2607. 19442v1 Announce Type: cross Abstract: Machine unlearning is commonly evaluated by matching a retrained oracle on trained probes.
UpgradeBench is a decision‑centric longitudinal benchmark that evaluates how fine‑tuned language‑model specialists should be handled when new base‑model releases occur. It covers four consecutive Qwen releases, a continuation checkpoint, six tasks, two model sizes, and OLMo checkpoints with known training lineage, and examines whether retraining, adapter transfer, or other recovery strategies improve specialist performance. The benchmark reveals that upgrade gains vary by task and release interval, that direct adapter copying is sensitive to pretraining distance, and that teacher relabeling can recover specialists without new annotations. "whyItMatters":"The study provides actionable insights into the cost‑effective management of specialist models across model releases, showing how to balance retraining effort with performance gains."
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:2608. 07719v1 Announce Type: new Abstract: Offline reinforcement learning repeatedly trains policies from a fixed transition pool, making redundant data costly across seeds and hyperparameters, while naive subsampling can remove rare transitions needed for long-horizon credit assignment.