Searcharxiv⌕ Search

arXiv subjects

Adam Belfki

Publications and source records attributed to Adam Belfki.

6 recordsLinked to original sources

How VLMs Plan What's Next

When a vision-language model describes an image, it names one object at a time and it rarely repeats itself, yet how the model plans this order is far from obvious. We characterize the mechanism a VLM uses to decide where to look before it looks. The mechanism has four main parts. First, binding heads match the relevant text tokens against the image and resolve them into pointers to the matched image region. Second, the same binding heads bind each written answer to the region it came from, and the model turns it into a progress signal: a cursor, the position in the image up to which the list has been written. Third, a select step chooses the nearest unwritten region after the cursor, or stops. Finally, gaze heads execute the chosen pointer and return a concept of what is drawn there, which language heads and the late MLPs turn into a word in the language of the question. We validate each part of this mechanism through causal patching in Qwen3-VL models on a synthetic visual retrieval task, and test our mechanisms against alternative hypotheses. We further show that the mechanism holds across input formats, natural photographs, and languages, and that its binding, relay, and gaze heads appear at every model size from 2B to 32B. This work allows us to investigate how a VLM turns a question and an image into a sequence of answers, layer by layer and token by token. More broadly, it opens a way to study how models plan over complex multi-step tasks with long rollouts. Our project is available at https://plan.baulab.info/

cs.CV↗

In-Context Learning Beyond Copying: A Training-Time Analysis of Abstractive ICL

The sudden emergence of induction circuits and the ability to perform in-context copying in LLMs coincides with a sharp drop in training loss over a short span of the training process. This co-occurrence has been cited as evidence that it marks the point at which a wide range of in-context learning (ICL) capabilities begin to emerge. To analyze the training dynamics of few-shot abstractive ICL capabilities (i.e., tasks where the answer is not contained in the input context), we propose Hapax, a training regime that omits the loss contribution of tokens predictable by the induction circuit to control training pressure for copying. Using this intervention, we find that abstractive ICL capabilities can still emerge, despite a substantial reduction in inductive copying and the absence of the characteristic phase changes observed during Vanilla training. Analysis of attention heads shows that Hapax-trained models develop fewer and weaker induction heads. Together, these results suggest that the developmental link between emergent phenomena associated with induction circuits and the emergence of abstractive ICL is weaker than previously hypothesized. More broadly, Hapax provides a controlled framework for analyzing how copying-related training signals shape the development of in-context learning.

cs.CL↗

Dyna-5G: Dynamic Role Switching for Self-Organizing 5G M2M Networks

M2M deployments such as drone swarms demand mission-critical communication: km-scale range, strong per-device identity and mutual authentication, and deterministic QoS for bandwidth-intensive payloads. Cellular 5G uniquely satisfies all of these, yet it has seen limited adoption in autonomous fleets. The barrier is not capability but resilience: today's 5G networks assume fixed infrastructure, and when the base station fails, recovery is uniquely complex. Unlike simpler wireless protocols where devices can transparently switch nodes, 5G failure requires reconstructing distributed state such as authentication contexts, QoS bindings, tunnels, and RRC state machines across the fleet, a process that no existing system automates. We present Dyna-5G, which makes this happen. Dyna-5G is the first 5G Standalone-compliant framework for dynamic role switching in M2M fleets, where any device can assume the role of 5G Core, RAN, or UE at runtime. It orchestrates failure detection, leader selection, and coordinated state teardown and re-establishment, all without modifying 3GPP protocols. We evaluate Dyna-5G on a high-fidelity software emulation testbed, with Open5GS and srsRAN, across hundreds of trials with up to 10 drones. Control-plane overhead averages 0.47 Mb/s (approximately 0.47% of a 100 Mb/s bearer), while failure recovery completes in about 2.5 s, of which approximately 86% is due to stack-dependent cellular procedures. Dyna-5G's orchestration logic itself adds only about 175 ms per reconfiguring role. All tested missions complete successfully, even under injected leader crashes.

cs.NI↗

Agents of Chaos

We report an exploratory red-teaming study of autonomous language-model-powered agents deployed in a live laboratory environment with persistent memory, email accounts, Discord access, file systems, and shell execution. Over a two-week period, twenty AI researchers interacted with the agents under benign and adversarial conditions. Focusing on failures emerging from the integration of language models with autonomy, tool use, and multi-party communication, we document eleven representative case studies. Observed behaviors include unauthorized compliance with non-owners, disclosure of sensitive information, execution of destructive system-level actions, denial-of-service conditions, uncontrolled resource consumption, identity spoofing vulnerabilities, cross-agent propagation of unsafe practices, and partial system takeover. In several cases, agents reported task completion while the underlying system state contradicted those reports. We also report on some of the failed attempts. Our findings establish the existence of security-, privacy-, and governance-relevant vulnerabilities in realistic deployment settings. These behaviors raise unresolved questions regarding accountability, delegated authority, and responsibility for downstream harms, and warrant urgent attention from legal scholars, policymakers, and researchers across disciplines. This report serves as an initial empirical contribution to that broader conversation.

cs.AI↗

MIB: A Mechanistic Interpretability Benchmark

How can we know whether new mechanistic interpretability methods achieve real improvements? In pursuit of lasting evaluation standards, we propose MIB, a Mechanistic Interpretability Benchmark, with two tracks spanning four tasks and five models. MIB favors methods that precisely and concisely recover relevant causal pathways or causal variables in neural language models. The circuit localization track compares methods that locate the model components - and connections between them - most important for performing a task (e.g., attribution patching or information flow routes). The causal variable localization track compares methods that featurize a hidden vector, e.g., sparse autoencoders (SAEs) or distributed alignment search (DAS), and align those features to a task-relevant causal variable. Using MIB, we find that attribution and mask optimization methods perform best on circuit localization. For causal variable localization, we find that the supervised DAS method performs best, while SAE features are not better than neurons, i.e., non-featurized hidden vectors. These findings illustrate that MIB enables meaningful comparisons, and increases our confidence that there has been real progress in the field.

cs.LG↗

NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internals

We introduce NNsight and NDIF, technologies that work in tandem to enable scientific study of the representations and computations learned by very large neural networks. NNsight is an open-source system that extends PyTorch to introduce deferred remote execution. The National Deep Inference Fabric (NDIF) is a scalable inference service that executes NNsight requests, allowing users to share GPU resources and pretrained models. These technologies are enabled by the Intervention Graph, an architecture developed to decouple experimental design from model runtime. Together, this framework provides transparent and efficient access to the internals of deep neural networks such as very large language models (LLMs) without imposing the cost or complexity of hosting customized models individually. We conduct a quantitative survey of the machine learning literature that reveals a growing gap in the study of the internals of large-scale AI. We demonstrate the design and use of our framework to address this gap by enabling a range of research methods on huge models. Finally, we conduct benchmarks to compare performance with previous approaches. Code, documentation, and tutorials are available at https://nnsight.net/.

cs.LG↗