ECLIPTICA is a framework that makes large language model alignment instruction‑driven and runtime‑controllable, allowing natural‑language alignment instructions to modulate behavior on the fly. It introduces Contrastive Instruction‑Tuned Alignment (CITA), which combines supervised fine‑tuning with contrastive preference optimization anchored to a reference model, ensuring stable policy updates within a shared neighborhood. The authors release an ECLIPTICA benchmark of 3,000 controlled cases and demonstrate that CITA achieves 86.7% instruction‑alignment efficiency on Llama‑3.1‑8B, outperforming DPO, GRPO, and PPO.
arXiv:2510. 07315v2 Announce Type: replace-cross Abstract: Large Language Models (LLMs) have catalyzed vibe coding, where users leverage LLMs to generate and iteratively refine code through natural language interactions until it passes their vibe check.
arXiv:2606. 22686v2 Announce Type: replace-cross Abstract: Modern Large Language Models (LLMs) rely on extensive safety alignment, yet the mechanistic basis of refusal remains opaque.
arXiv:2606. 00145v1 Announce Type: cross Abstract: Vision-language-action (VLA) agents can execute natural-language instructions, yet deployed systems still lack an operational interface: deciding when the instruction is complete.
arXiv:2608. 02639v1 Announce Type: cross Abstract: Production prompts rarely carry a single instruction.
EOPSA (Efficient On-Policy Self-Distilled Safety Alignment) addresses inefficiencies in On-Policy Self-Distillation (OPSD) for safety alignment by focusing training on safety-critical tokens. It introduces Adaptive Rollout Scheduling, which limits generation length based on a Teacher Rescue Rate metric, and Selective Distillation, which filters out safety-neutral tokens to concentrate gradient updates on safety-pivotal transitions. Experiments on models up to 32B parameters show that EOPSA reduces rollout computation by about 50% and backpropagates through only roughly 2% of tokens, outperforming full-token distillation baselines in safety compliance and reasoning retention.
arXiv:2606. 28998v1 Announce Type: cross Abstract: Large Language Model (LLM) alignment trains an LLM using preference data to produce outputs that better meet established quality standards.
Qwen-MT-Lumen is an open‑source translation model submitted to the WMT 2026 General Machine Translation Shared Task. It employs a four‑stage pipeline that combines continued pre‑training, supervised fine‑tuning, online preference‑optimization with an LLM judge, and on‑policy distillation from a larger teacher to a Qwen3‑14B student. The system also uses structure‑aware inference and selective repair to preserve formatting and correct decoding errors, achieving an overall mean score of 4.60 on a 1–5 rubric and matching leading proprietary systems in both translation quality and instruction adherence.
The paper introduces Many-Tier Instruction Hierarchy (ManyIH), a new framework for resolving conflicts among instructions with arbitrarily many privilege levels in large language model agents. It presents ManyIH-Bench, a benchmark featuring 853 agentic tasks that require navigating up to 12 levels of conflicting instructions across 46 real-world agents. Experiments show current models achieve only about 40% accuracy when instruction conflict scales, highlighting a gap in fine-grained, scalable conflict resolution.
arXiv:2605.28360v3 Announce Type: replace Abstract: Automatic prompt optimization (APO) has driven significant gains in LLM-based agentic workflows. However, most existing methods treat each task's p...
The paper introduces ELCD, a latent conflict detector that verifies LLM outputs after generation to catch instruction conflicts that static input checks miss. ELCD builds a hidden-state representation from the final-token embedding and the mean-pooled response embedding, then trains a pairwise margin ranking objective to distinguish compliant from drifting responses. Experiments on five large language models show ELCD outperforms baselines, boosting PR-AUC for Llama‑2‑7B by ~30 percentage points and cutting FPR95 for Mistral‑7B to 2.67%.
arXiv:2511. 04694v5 Announce Type: replace-cross Abstract: As large language model (LLM) based systems take on high-stakes roles in real-world decision-making, they must reconcile competing instructions from multiple sources within a single prompt context.
The paper introduces XIH-Bench, a multilingual benchmark for evaluating instruction hierarchy (IH) compliance in large language models across six languages, four domains, and three IH settings. It finds that IH compliance varies with language, showing an asymmetry where a language can strengthen compliance in a higher-priority role but disrupt it in a lower-priority role, and that cross-language conflicts yield higher compliance than same-language conflicts—a phenomenon called the Language Boundary Effect. The study also highlights that language specialization can make lower-priority instructions in favored languages harder to override, raising potential reliability and security concerns.