arXiv:2609.24932v1 Announce Type: new Abstract: Despite the success of neural models in natural language processing, their black-box nature limits interpretability and conceals the linguistic phenome...
The paper introduces an inductive method for learning concept embeddings from cross‑linguistic colexification networks combined with short textual descriptions. Unlike previous transductive approaches, this framework can embed unseen concepts using only a brief definition. The resulting embeddings correlate with cross‑linguistic similarity measures and generalize well to new concepts.
arXiv:2608.29529v1 Announce Type: cross Abstract: Semantic alignment between specialized normative texts is challenging when equivalent requirements use different terms, syntax, and levels of abstrac...
The paper extends mechanistic interpretability of large language models by modeling concepts as low‑dimensional non‑linear manifolds rather than linear subspaces. It introduces a concept‑based alignment (CBA) score to compare these manifolds across layers and models, revealing block structures in intermediate layers, a shift from syntax‑dominated to mixed syntactic‑semantic concepts, and training‑dependent multilingual sharing. The study also shows that alignment patterns differ across model families and training stages, with adjacent stages aligning more closely than distant ones.
The paper introduces a comparative explainability framework for auditing DeBERTa‑v3 in zero‑shot medical abstract classification. It evaluates five explanation methods—SHAP, LIME, occlusion, Input × Gradient, and Attention × Gradient—using a natural language inference engine on a balanced corpus of 1,000 abstracts per diagnostic category. The study finds that explanatory stability aligns with predictive certainty, identifies three systemic failure mechanisms, and recommends combining multiple explanation methods and quantitative agreement metrics for transformer‑based medical text classifiers.
arXiv:2606. 08497v1 Announce Type: new Abstract: As deep language models (DLMs) are increasingly deployed in high-stakes domains such as healthcare, understanding their decision rationale becomes paramount for ensuring trust, safety, and accountability.
arXiv:2608. 06417v1 Announce Type: new Abstract: The proliferation of misinformation online has driven demand for scalable detection systems.
arXiv:2605. 05103v3 Announce Type: replace-cross Abstract: We introduce the \textbf{Concept Field} of a text corpus: a local drift field with pointwise uncertainty, estimated in sentence-embedding space from the deltas between consecutive sentences.
arXiv:2607.00171v2 Announce Type: replace Abstract: Text embeddings are standard for semantic similarity tasks, yet their evaluation remains an open challenge. Current benchmarks are static, cover on...
The paper introduces the Concept Field, a local drift field with pointwise uncertainty derived from sentence-embedding deltas in a text corpus. It scores candidate sentence transitions using the mean absolute z-distance (ζ) to quantify groundedness, offering a model‑agnostic, corpus‑attributable metric. The authors implement a Vector Sequence Database to store embeddings and delta metadata, and demonstrate the method’s effectiveness in detecting groundedness in the U.S. Code of Federal Regulations and novelty in Project Gutenberg, achieving strong selective classification performance under a triage policy.
arXiv:2608.27813v1 Announce Type: new Abstract: Structural probes were introduced by Hewitt and Manning to reconstruct syntactic trees from a neural language model's latent representations. They are...
arXiv:2606. 29180v1 Announce Type: new Abstract: A Knowledge Graph (KG) represents facts as structured triples and is widely used to organize relational knowledge across diverse domains.