The study investigates how information about words and passage context is represented across the layers of the Llama‑3.1‑8B language model during silent reading, using EEG recordings. Decoders trained on each layer show that word‑level decodability peaks in early layers and diminishes toward the final layer, while a context‑tracking component increases with depth. Even a single 0.5‑second EEG window can capture passage‑level information, and control experiments confirm that this context tracking is driven by the text rather than by word decoding or session artifacts.
arXiv:2608. 20186v1 Announce Type: new Abstract: Non-invasive decoding of inner speech faces a fundamental data problem: a corpus pairing brain activity with a person's spontaneous inner monologue cannot be collected, and the available proxy paradigms (cued repetitive and retrospectively reported generative inner speech) are slow to acquire, poorly time-locked, and subject compliance is unverifiable.
arXiv:2604. 16370v2 Announce Type: replace-cross Abstract: Decoding natural language from non-invasive electroencephalography (EEG) remains constrained by low signal-to-noise ratio and limited information bandwidth.
arXiv:2606. 26519v2 Announce Type: replace Abstract: Large language models (LLMs) can make scientific software easier to use.
arXiv:2606. 26519v1 Announce Type: new Abstract: Large language models (LLMs) can make scientific software easier to use.
The study investigates the internal workings of an audio language model (Qwen3-Omni) by applying a logit lens to its middle layers. It finds that the model’s reasoning about spoken questions becomes legible in words before any token is emitted, revealing language‑agnostic, paralinguistic, and temporally distinct signals that are causally used in the network’s decision process. The authors demonstrate that these signals can be isolated and mapped to specific layers, providing a qualitative account of how the model processes audio input.
arXiv:2606. 01923v1 Announce Type: cross Abstract: Large Language Models (LLMs) frequently exhibit "contextual disregard" when faced with input evidence that conflicts with their internal parametric memory, leading to persistent factual hallucinations.
The paper introduces a three-level evaluation framework—behavioral deployment, LM-head readout, and probe recoverability—to distinguish whether a language model fails a syntactic test by not encoding structure or by failing to use it. Using a trilingual control-dependency benchmark, the authors find that probe recoverability consistently exceeds LM-head readout, which in turn exceeds behavioral deployment across seven models and three languages, with the largest gap observed in Qwen3-0.6B Instruct. Layer-localized activation patching shows that instruction tuning shifts the decoded layer later, suggesting decoding favors surface shortcuts and that behavioral evaluation understates what models encode while probing alone overstates what they deploy.
Neural State Prediction (NSP) is a latent‑predictive framework designed to curb shortcut learning in EEG foundation models. By using a target encoder updated with an exponential moving average, identity residualization, and topology‑separated context, NSP constrains both the prediction target and the available context. Trained on 2.2 million EEG segments, NSP outperforms baselines on 14 datasets in the EEG‑FM‑Bench, achieving 63.94 % macro balanced accuracy.
arXiv:2608. 04156v1 Announce Type: new Abstract: Electroencephalography (EEG) analysis extends beyond assigning predefined labels to recordings; it requires workflows connecting natural-language instructions, signal processing, quantitative evidence, and scientific interpretation.
arXiv:2608. 01481v1 Announce Type: new Abstract: Short segments of perceived speech can be retrieved from non-invasive magnetoencephalographic (MEG) recordings by deep networks trained with a CLIP-style objective against wav2vec 2.
arXiv:2607. 20720v1 Announce Type: cross Abstract: Harnessing the potential of electroencephalography (EEG) for brain research is fundamentally limited by intrinsic noise and the diffuse projection of brain-generated activity over the scalp.
arXiv:2609.24095v1 Announce Type: new Abstract: While deep learning has enabled language decoding from intracranial brain recordings, extending this capability to non-invasive recordings remains an u...