Novel metaphor comprehension involves complex semantic processes and linguistic creativity, making it an interesting task for studying language models (LMs). This study investigates whether surprisal, a probabilistic measure of predictability in LMs, correlates with annotations of metaphor novelty in different datasets. We analyse the surprisal of metaphoric words in corpus-based and synthetic metaphor datasets using 16 causal LM variants. We propose a cloze-style surprisal method that conditions on full-sentence context. Results show that LM surprisal yields significant moderate correlations with scores/labels of metaphor novelty. We further identify divergent scaling patterns: on corpus-based data, correlation strength decreases with model size (inverse scaling effect), whereas on synthetic data it increases (quality–power hypothesis). We conclude that while surprisal can partially account for annotations of metaphor novelty, it remains limited as a metric of linguistic creativity. Code and data are publicly available: https://github.com/OmarMomen14/surprisal-metaphor-novelty

We investigate whether pre-training exclusively on dialogue data results in formally and functionally apt small language models. Based on this pre-trained llamalogue model, we employ a variety of fine-tuning strategies to enforce “more communicative” text generations by our models. Although our models underperform on most standard BabyLM benchmarks, they excel at dialogue continuation prediction in a minimal pair setting. While PPO fine-tuning has mixed to adversarial effects on our models, DPO fine-tuning further improves their performance on our custom dialogue benchmark.

Analysing texts spanning long periods of time is critical for researchers in historical linguistics and related disciplines. However, publicly available corpora suitable for such analyses are scarce. The Project Gutenberg (PG) corpus presents a significant yet underutilized opportunity in this context, due to the absence of accurate temporal metadata. We take advantage of language models and information retrieval to explore four sources of information – Open Web, Wikipedia, Open Library API, and PG books texts – to add missing temporal metadata to the PG corpus. Through 20 experiments employing state-of-the-art Large Language Models (LLMs) and Retrieval-Augmented Generation (RAG) methods, we estimate the production years of all PG books. We curate an enriched metadata repository for the PG corpus and propose a refined version for it, which includes 53,774 books with a total of 3.8 billion tokens in 11 languages, produced between 1600 and 2000. This work provides a new resource for computational linguistics and humanities studies focusing on diachronic analyses. The final dataset and all experiments data are publicly available (https://github.com/OmarMomen14/pg-dates).

Descriptions are a central component of literary texts, yet their systematic identification remains a challenge. This work suggests an approach to identifying sentences describing spatial conditions in literary text. It was developed iteratively on German literary text and extended to non-literary text to evaluate its applicability across textual domains. To assess the robustness of the method, we involved both humans and a selection of state-of-the-art Large Language Models (LLMs) in annotating a collection of sentences regarding their descriptiveness and spatiality. We compare the annotations across human annotators and between humans and LLMs. The main contributions of this paper are: (1) a set of annotation guidelines for identifying spatial descriptions in literary texts, (2) a curated dataset of almost 4,700 annotated sentences of which around 500 are spatial descriptions, produced through in-depth discussion and consensus among annotators, and (3) a pilot study of automating the task of spatial description annotation of German texts. We publish the codes and all human and LLM annotations for the public to be used for research purposes only.

Text simplification is an intralingual translation task in which documents, or sentences of a complex source text are simplified for a target audience. The success of automatic text simplification systems is highly dependent on the quality of parallel data used for training and evaluation. To advance sentence simplification and document simplification in German, this paper presents DEplain, a new dataset of parallel, professionally written and manually aligned simplifications in plain German “plain DE” or in German: “Einfache Sprache”. DEplain consists of a news-domain (approx. 500 document pairs, approx. 13k sentence pairs) and a web-domain corpus (approx. 150 aligned documents, approx. 2k aligned sentence pairs). In addition, we are building a web harvester and experimenting with automatic alignment methods to facilitate the integration of non-aligned and to be-published parallel documents. Using this approach, we are dynamically increasing the web-domain corpus, so it is currently extended to approx. 750 document pairs and approx. 3.5k aligned sentence pairs. We show that using DEplain to train a transformer-based seq2seq text simplification model can achieve promising results. We make available the corpus, the adapted alignment methods for German, the web harvester and the trained models here: https://github.com/rstodden/DEPlain.