Frames LLM routing as a sequential decision process and introduces LLMRouter plus the xRouteBench benchmark to develop, evaluate, and deploy learned routing policies across heterogeneous LLMs, optimizing response quality versus inference cost.
Evaluates whether LLM-driven storyteller agents preserve long-horizon logical consistency under adversarial player interventions. Introduces NCP-Bench (100 movie-derived narrative environments) with structured trajectory/commitments and automatic violation checks; finds strong LLMs often contradict themselves across multi-turn interactions.
A MoE causal large language model for long-horizon agents, coding, and multi-step reasoning: 2.4T parameters (95B activated), native 262,144-token context (extensible to 1,010,000), multi-token prediction, and configurable thinking-mode reasoning controls.
Human-annotated text dataset that labels perceived “AI slop” with a continuous human slop_score (-1 / 0 / +1) plus provenance metadata (source_dataset, source_row_id, content_hash). Collected via Bench Labs SlopFinder from public datasets for training classifiers and studying subjective perception.
Provides a curated benchmark of 170 real-world, multilingual code-refactoring instances to evaluate AI coding agents on large-scale, behavior-preserving, cross-file refactors. Each task includes rewritten issue descriptions and manually reviewed test suites to avoid over- and under-constraining evaluations.
A research report proposing a continual-learning agent workflow that pairs recursive self-improvement with a Mixture-of-LoRA design: freeze a foundation model, compose specialist LoRA adapters routed per user turn, and support them with long-context RL and post-training infrastructure.
Provides a year-scale multimodal benchmark and evaluation framework for on-device long-term memory in personal assistants, built from real mobile user trajectories. Tests memory construction, retrieval, updating, temporal reasoning, and implicit preference inference, and includes a knowledge-grounded synthesis pipeline to form coherent long-horizon trajectories.
Encodes videos into a Film Knowledge Graph and reconstructs them to learn agent-native, editable video representations for agentic reasoning and manipulation. Uses agentic auto-encoding with dual-loop textual-gradient optimization, reports large reconstruction gains, and releases a benchmark and dataset.
Uses a stronger 'builder' model at inference time to construct executable harnesses that boost weaker target models without parameter updates, mainly by turning unstable reasoning into deterministic code, routing, and strict answer-format enforcement.
Externalizes persistent scene state into a camera-indexed world bank and designs a long-horizon teacher whose sparse-attention supervision is distilled into a three-step student, enabling responsive, low-latency interactive long-horizon video generation with bounded denoiser context.
Turns natural-language instructions into end-to-end Android automation that drives real devices and captures Logcat and screenshots. Uses a multimodal “Dynamic‑First, Coordinate‑Fallback” locator and a native MCP server for IDE integration, supporting fast Flash (3–5s/step) and verifiable Pro workflows.
Systematically evaluates LLM-driven autonomous agents on long-horizon AI research tasks using rule-based within-run metrics (Solution Framing, Execution, Feedback Control). Focuses on experience reuse and harness effects across 36 tasks and seven frontier models, finding agents act more like engineering optimizers than autonomous researchers.