Targeted Visualization of the Backbone of Encoder LLMs
Roberts I, Schulz A, Schr{\ S, Hermes L, Hammer B (2026)
In: Machine Learning and Principles and Practice of Knowledge Discovery in Databases. Cerrato M, Kalinauskait{\.{e}} D{\.{e}}, Luko{\v{s}}evi{\v{c}}ius M, Pechenizkiy M, v{S}}utien{\.{e}} K (Eds); Cham: Springer Nature Switzerland: 82-97.
Konferenzbeitrag | Englisch
Download
Es wurden keine Dateien hochgeladen. Nur Publikationsnachweis!
Autor*in
Roberts, Isaac;
Schulz, Alexander;
Schr{\, Sarah;
Hermes, Luca;
Hammer, Barbara
Herausgeber*in
Cerrato, Mattia;
Kalinauskait{\.{e}}, Danguol{\.{e}};
Luko{\v{s}}evi{\v{c}}ius, Mantas;
Pechenizkiy, Mykola;
v{S}}utien{\.{e}}, Kristina
Abstract / Bemerkung
Attention-based Large Language Models (LLMs) are the state-of-the-art in natural language processing (NLP). The two most common architectures are encoders such as BERT, and decoders like the GPT models. Despite the success of encoder models, on which we focus in this work, they also bear several risks, including issues with bias or their susceptibility for adversarial attacks, signifying the necessity for explainable AI to detect such issues. While there do exist various local explainability methods focusing on the prediction of single inputs, global methods based on dimensionality reduction for classification inspection, which have emerged in other domains and that go further than just using t-SNE in the embedding space, are not widely spread in NLP.
Erscheinungsjahr
2026
Titel des Konferenzbandes
Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Seite(n)
82-97
ISBN
978-3-032-25305-7
Page URI
https://pub.uni-bielefeld.de/record/3016844
Zitieren
Roberts I, Schulz A, Schr{\ S, Hermes L, Hammer B. Targeted Visualization of the Backbone of Encoder LLMs. In: Cerrato M, Kalinauskait{\.{e}} D{\.{e}}, Luko{\v{s}}evi{\v{c}}ius M, Pechenizkiy M, v{S}}utien{\.{e}} K, eds. Machine Learning and Principles and Practice of Knowledge Discovery in Databases. Cham: Springer Nature Switzerland; 2026: 82-97.
Roberts, I., Schulz, A., Schr{\, S., Hermes, L., & Hammer, B. (2026). Targeted Visualization of the Backbone of Encoder LLMs. In M. Cerrato, D. {\.{e}} Kalinauskait{\.{e}}, M. Luko{\v{s}}evi{\v{c}}ius, M. Pechenizkiy, & K. v{S}}utien{\.{e}} (Eds.), Machine Learning and Principles and Practice of Knowledge Discovery in Databases (pp. 82-97). Cham: Springer Nature Switzerland.
Roberts, Isaac, Schulz, Alexander, Schr{\, Sarah, Hermes, Luca, and Hammer, Barbara. 2026. “Targeted Visualization of the Backbone of Encoder LLMs”. In Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ed. Mattia Cerrato, Danguol{\.{e}} Kalinauskait{\.{e}}, Mantas Luko{\v{s}}evi{\v{c}}ius, Mykola Pechenizkiy, and Kristina v{S}}utien{\.{e}}, 82-97. Cham: Springer Nature Switzerland.
Roberts, I., Schulz, A., Schr{\, S., Hermes, L., and Hammer, B. (2026). “Targeted Visualization of the Backbone of Encoder LLMs” in Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Cerrato, M., Kalinauskait{\.{e}}, D. {\.{e}}, Luko{\v{s}}evi{\v{c}}ius, M., Pechenizkiy, M., and v{S}}utien{\.{e}}, K. eds. (Cham: Springer Nature Switzerland), 82-97.
Roberts, I., et al., 2026. Targeted Visualization of the Backbone of Encoder LLMs. In M. Cerrato, et al., eds. Machine Learning and Principles and Practice of Knowledge Discovery in Databases. Cham: Springer Nature Switzerland, pp. 82-97.
I. Roberts, et al., “Targeted Visualization of the Backbone of Encoder LLMs”, Machine Learning and Principles and Practice of Knowledge Discovery in Databases, M. Cerrato, et al., eds., Cham: Springer Nature Switzerland, 2026, pp.82-97.
Roberts, I., Schulz, A., Schr{\, S., Hermes, L., Hammer, B.: Targeted Visualization of the Backbone of Encoder LLMs. In: Cerrato, M., Kalinauskait{\.{e}}, D.{\.{e}}, Luko{\v{s}}evi{\v{c}}ius, M., Pechenizkiy, M., and v{S}}utien{\.{e}}, K. (eds.) Machine Learning and Principles and Practice of Knowledge Discovery in Databases. p. 82-97. Springer Nature Switzerland, Cham (2026).
Roberts, Isaac, Schulz, Alexander, Schr{\, Sarah, Hermes, Luca, and Hammer, Barbara. “Targeted Visualization of the Backbone of Encoder LLMs”. Machine Learning and Principles and Practice of Knowledge Discovery in Databases. Ed. Mattia Cerrato, Danguol{\.{e}} Kalinauskait{\.{e}}, Mantas Luko{\v{s}}evi{\v{c}}ius, Mykola Pechenizkiy, and Kristina v{S}}utien{\.{e}}. Cham: Springer Nature Switzerland, 2026. 82-97.
Google Scholar
ISBN Suche