The paper demonstrates that a hypernetwork can generate tiny specialist models for individual transformations in the ARC-1D benchmark directly from a few demonstrations. These generated parameters form a structured weight space, enabling partial compositional generalisation and improved performance on unseen transformations when explicit task identifiers are omitted. The study provides a proof of concept that few‑shot task context can be compiled on‑the‑fly into compact executable model parameters that support reuse and generalisation beyond known functions.
arXiv:2607. 06764v1 Announce Type: new Abstract: Recent progress on ARC-AGI-1 from disclosed architectures has come broadly from two regimes: heavy test-time compute over frontier models (evolutionary search, exhaustive sampling, extended chain-of-thought), or benchmark-specific training in which small models are fine-tuned on ARC data, often with task-specialized architectures.
arXiv:2607. 19604v1 Announce Type: cross Abstract: Injecting factual knowledge into large language models (LLMs) reliably and at scale remains an open challenge.
arXiv:2606. 30704v1 Announce Type: cross Abstract: Large language models (LLMs) excel across a wide range of tasks, yet their instance-specific solutions often lack the structural consistency needed for reliable deployment.
arXiv:2610.34413v1 Announce Type: cross Abstract: Large audio language models (LALMs) have achieved strong performance on standard speech and audio understanding tasks, yet their ability to systemati...
The paper introduces a two‑step test‑time training protocol, Embed‑TTT, for Vision ARC (VARC) that first fine‑tunes only the task embedding and then fine‑tunes the backbone. This approach consistently produces task embeddings that better align with the underlying rules, improves retrieval and linear probing, and recovers the geometric structure of parametric rules. Even fine‑tuning only the tiny embedding component solves a significant portion of ARC‑AGI‑1, ConceptARC, and Mini‑ARC tasks, while the full two‑step pipeline further enhances performance and demonstrates compositional rule interpolation.
arXiv:2610.01054v1 Announce Type: cross Abstract: In-context learning (ICL) enables language models to perform new tasks from demonstrations without weight updates. However, every ICL inference requi...
MESA is an evaluation framework that precisely measures language‑conditioned policy generalization in robotics. It offers five evaluation suites that separate in‑distribution performance from four axes of generalization—spatial configuration, object instance, object category, and subtask composition—allowing researchers to pinpoint why policies fail. The authors also introduce MESA‑Gen, a task and demonstration generation pipeline, and release two benchmark suites: a single‑arm MESA and BiMESA for bimanual generalization, with baseline results showing current models achieve only about 50% success, indicating significant room for improvement.
The paper introduces w$^{2}$VLA, a Vision‑Language‑Action model that restructures information flow to separate declarative knowledge (concepts and entity semantics) from procedural knowledge (how to act). By modulating the robot state sequence with visual, spatial, and skill information in a compositional, interpretable way, the model achieves robust behavior cloning and unprecedented zero‑shot skill transfer to unseen objects, overcoming brittleness seen in current state‑of‑the‑art VLA models.
arXiv:2602. 12952v3 Announce Type: replace Abstract: Adapting large pre-trained models to downstream tasks often produces task-specific parameter updates that are expensive to relearn for every model variant.
arXiv:2610.40934v1 Announce Type: new Abstract: In this work, we provide a formal framework and mathematical definition for the subset of foundation models focused on enabling the property of "any-ta...
arXiv:2506. 11042v2 Announce Type: replace Abstract: Parameter-efficient fine-tuning (PEFT) has emerged as a resource-efficient strategy for adapting Pretrained Foundation Models (PFMs) by learning a small number of task-specific updates $\Delta W$.
arXiv:2606.30015v2 Announce Type: replace Abstract: For LLMs, skills are textual recipes encoding mature problem-solving experience. Despite widespread deployment, their utility is limited by the mod...