arXiv:2606. 07532v1 Announce Type: cross Abstract: RLHF-trained models are systematically biased toward agreement over accuracy, a structural property of the training process.
arXiv:2606. 07532v2 Announce Type: replace-cross Abstract: RLHF-trained models are systematically biased toward agreement over accuracy, a structural property of the training process.
arXiv:2606. 13197v1 Announce Type: new Abstract: Multi-agent debate (MAD) can improve large language model reasoning, but fixed debate pipelines often waste computation and can amplify correlated errors among similar agents.
arXiv:2606. 10475v1 Announce Type: cross Abstract: Multi-agent debate frameworks have been shown to improve large language model performance in convergent tasks, but they are currently optimized in a way that heavily favors final output accuracy rather than stability of the process.
arXiv:2606. 24976v1 Announce Type: cross Abstract: Foundation-model agents in multi-step, open-ended environments frequently suffer from compounding errors, where early mistakes contaminate long-horizon trajectories.
arXiv:2601. 19921v2 Announce Type: replace-cross Abstract: Multi-agent debate (MAD) is widely used to improve large language model (LLM) performance through test-time scaling, yet recent work shows that vanilla MAD often underperforms simple majority vote despite higher computational cost.
arXiv:2605. 12991v3 Announce Type: replace-cross Abstract: LLM-based multi-agent pipelines flip from correct to incorrect answers under simulated peer disagreement at rates we term yield, a vulnerability widely attributed to RLHF-induced sycophancy.
arXiv:2608. 13069v1 Announce Type: new Abstract: Large language models (LLMs) are predominantly aligned to function as passive, sycophantic assistants.
arXiv:2608. 17776v1 Announce Type: new Abstract: We demonstrate that RL finetuning an LLM using debate, a two-player adversarial game between a generator and a critic adjudicated by a weaker LLM judge, reduces reward hacking compared to a reinforcement learning from AI feedback (RLAIF) baseline.
arXiv:2606. 07897v1 Announce Type: new Abstract: Current AI models frequently exhibit epistemic sycophancy, endorsing claims to agree with a user.
arXiv:2608. 08210v1 Announce Type: new Abstract: Collaborative dialogue can end with apparent agreement while participants still differ on goals, assumptions, or execution plans, creating an \textbf{illusion of alignment (IoA)}.
Large language models (LLMs) are predominantly aligned to function as passive, sycophantic assistants. We challenge this default paradigm by empirically evaluating the cognitive plasticity of open-weight architectures when subjected to rigorous behavioral reprogramming.
Large language models achieve strong reasoning performance, but often at prohibitive training cost - a challenge that is especially acute for compact models ($\leq 4 \, \mathrm{B}$ parameters) trained under limited budgets. We introduce MADA-RL, a post-training framework that specializes compact models into generator and critic roles and trains them with a debate-aware learning signal, fine-tuning only a small subset of parameters via LoRA adapters.