United States

Evaluating and optimizing authorial style in long-form story generation is challenging because style judgments often rely on subjective human voting, and there is no stable automatic evaluation method. We propose a two-stage pipeline. First, we train a style-similarity judge by fine-tuning a sentence-transformer with authorship-verification supervision, and calibrate its similarity outputs into a bounded [0,1] reward. Second, we use this judge as the primary reward in Group Relative Policy Optimization (GRPO) to fine-tune an 8B story generator for style-conditioned writing, avoiding the accept/reject supervision required by Direct Preference Optimization (DPO). Across four target authors (Mark Twain, Jane Austen, Charles Dickens, Thomas Hardy), the GRPO-trained 8B model achieves higher style scores than open-weight baselines, with an average style score of 0.893 across authors. These results suggest that AV-calibrated reward modeling provides a practical mechanism for controllable long-form style transfer under moderate model size and training budget.

ACL 2026

Capturing Classic Authorial Style in Long-Form Story Generation with GRPO Fine-Tuning

literary nlp

long-form text generation

natural language generation

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/acl2026

Registration Required

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

workshop paper

The thematic fit estimation task measures semantic arguments’ compatibility with a specific semantic role for a specific predicate. We investigate if LLMs have consistent, expressible knowledge of event arguments’ thematic fit by experimenting with various prompt designs, manipulating input context, reasoning, and output forms. We set a new state-of-the-art on thematic fit benchmarks, but show that closed and open weight LLMs respond differently to our prompting strategies: Closed models achieve better scores overall and benefit from multi-step reasoning, but they perform worse at filtering out generated sentences incompatible with the specified predicate, role, and argument.

Uncovering Autoregressive LLM Knowledge of Thematic Fit in Event Representation

Recent work has shown that language models (LMs) trained on synthetic corpora can exhibit typological preferences that resemble cross-linguistic regularities in human languages, particularly for syntactic phenomena such as word order. In this paper, we extend this paradigm to differential argument marking (DAM), a semantic licensing system in which morphological marking depends on semantic prominence. Using a controlled synthetic learning method, we train GPT-2 models on 18 corpora implementing distinct DAM systems and evaluate their generalization using minimal pairs. Our results reveal a dissociation between two typological dimensions of DAM. Models reliably exhibit human-like preferences for natural markedness direction, favoring systems in which overt marking targets semantically atypical arguments. In contrast, models do not reproduce the strong object preference in human languages, in which overt marking in DAM more often targets objects rather than subjects. These findings suggest that different typological tendencies may arise from distinct underlying sources.

Differences in Typological Alignment in Language Models' Treatment of Differential Argument Marking

CHILDES is a central resource for language acquisition studies, yet computational tools for analyzing its syntactic structure remain limited. Leveraging the recent release of the UD-English-CHILDES treebank with gold-standard Universal Dependencies (UD) annotations, we train a state-of-the-art dependency parser specifically tailored to CHILDES. The parser more accurately captures syntactic patterns in child--adult interactions, outperforming widely used off-the-shelf English parsers, including SpaCy and Stanza.
Alongside the parser, we also release a Part-of-Speech tagger and an utterance-level construction tagger, which together form the open-source Syntactic Annotation Toolkit for Child--Adult InTeractions (CAIT). Through a detailed error analysis and a case study tracking the distribution of syntactic constructions across developmental time in CHILDES, we demonstrate the practical utility of the toolkit for large-scale, reproducible research on language acquisition.

A Syntactic Annotation Toolkit for Child–Adult InTeractions (CAIT)

Understanding how neural models represent human-interpretable concepts is challenging. Prior work has explored linear concept subspaces from diverse perspectives, such as probing and concept erasure. We introduce a unified framework to study these subspaces along two axes: $\textit{containment}$, which tests if a concept is fully represented in a subspace but not outside it, and $\textit{disentanglement}$, which tests for isolation from other concepts. In experiments on both text and speech models, we first highlight that concept subspaces may not be uniquely determined, and discuss the implications for concept subspace analysis. Then, we compare properties of concept subspaces estimated using five estimators, proposed in different communities. We find that (1) the choice of estimator impacts the containment and disentanglement properties; (2) the state-of-the-art concept erasure method, LEACE, performs well on both testing axes, but still struggles to generalize to unseen data; and (3) in HuBERT speech representations, phone information is both contained and disentangled from speaker information, while speaker information is hard to contain in a compact subspace, despite being disentangled from phones.

A framework for analyzing concept representations in neural models

This paper compares the empathetic quality of responses generated by humans and large language models (LLMs). We evaluate four LLMs that were widely used at the time of study—GPT-4, LLaMA-2-70B-Chat, Gemini-1.0-Pro, and Mixtral-8×7B-Instruct—against a human baseline using a large-scale between-subjects study. A total of 1,000 human participants evaluated the empathetic quality of human- and LLM-generated responses to 2,000 dialogue prompts spanning 32 positive and negative emotions. To complement human judgments, we also employed an LLM-as-judge (GPT-4o-mini) to assess the same responses. Across emotions and evaluators, LLM-generated responses were rated as significantly more empathetic than human-written responses. We also observed that both human judges and the LLM-as-judge tended to rate responses generated by their own group more favorably, indicating self-favoring tendencies. These findings highlight both the strong performance of contemporary LLMs in empathetic responding and the need to interpret human- and LLM-based evaluations with care.

Who Generates More Empathetic Responses—Humans or LLMs? A Comparative Evaluation with Human and LLM Judges

Language model training and inference ignore a fundamental linguistic fact -- there is a dependence between multiple sequences of text written by the same person. Prior work has shown that addressing this form of \textit{ecological fallacy} can greatly improve the performance of multiple smaller (~124M) GPT-based models. In this work, we ask if addressing the ecological fallacy by modeling the author's language context with a specific LM task (called HuLM) can provide similar benefits for a larger-scale model, an 8B Llama model. To this end, we explore variants that process an author's language in the context of their other temporally ordered texts. We study the effect of pre-training with this author context using the HuLM objective, as well as using it during fine-tuning with author context (\textit{HuFT:Human-aware Fine-Tuning}). Empirical comparisons show that addressing the ecological fallacy during fine-tuning alone using QLoRA improves the performance of the larger 8B model over standard fine-tuning. Additionally, QLoRA-based continued HuLM pre-training results in a human-aware model generalizable for improved performance over eight downstream tasks with linear task classifier training alone. These results indicate the utility and importance of modeling language in the context of its original generators, the authors.

Addressing the Ecological Fallacy in Larger LMs with Human Context

Attention sinks—initial tokens that absorb disproportionate attention 
mass—are widely regarded as softmax artifacts. We show they carry a 
second function: relaying morphosyntactic features for agreement 
computation. Masking sink positions across 67 BLiMP subtasks in five 
models (three architectures) produces a selective dissociation: 
agreement collapses (Δ = −0.50) while binding, NPI licensing, and 
island constraints degrade moderately (Δ = −0.17; d = −1.89). 
Activation patching isolates a write–propagate–read circuit: L0 
attention writes subject number to sinks, intermediate MLPs maintain 
it, and cliff-layer heads consume it—early-layer restoration recovers 
77% of the effect at 1B. The read phase is sparse at 1B (3 head 
indices) and distributed at 2.8B (32 heads), yet agreement-selective 
at both scales. The relay emerges at training steps 512–2000, 
coinciding with sink formation, and attenuates beyond ~25 tokens as 
direct content-to-content pathways become available. German 
replicates both behavioral and causal signatures; Korean replicates 
the behavioral pattern. These results establish sinks as positional 
relays: emergent fixed-address loci through which transformers route 
grammatical features across structurally intervening material.

Attention Sinks as Agreement Relays: Selective Collapse of Morphosyntactic Dependencies under Sink Token Removal

In what ways might statistical signals in linguistic input assist with the acquisition of syntax? Here we hypothesize a mechanism called collocational bootstrapping, in which regularities in word co-occurrence patterns can provide cues to syntactic dependencies. We investigate whether this mechanism can support the acquisition of English subject-verb agreement. First, we simulate language acquisition by training neural networks on synthetic datasets that vary in how predictable their subject-verb pairings are. We find that there is a range of variability levels at which these statistical learners robustly learn subject-verb agreement. We then analyze the variability of subject-verb pairings in child-directed language, and we find that the variability in such data falls within the range that supported robust generalization in our computational simulations. Taken together, these results suggest that collocational bootstrapping is a viable learning strategy for the type of input that children receive.

Collocational boostrapping: A hypothesis about the learning of subject-verb agreement in humans and neural networks

What can prompting an LLM teach us about the contours of a semantic category? While LLM embeddings are often regarded as high-quality semantic representations, it is not clear that prompting an LLM is a useful way to obtaining accurate extensional judgments (e.g., whether a DIY gun kit qualifies as a “firearm”). Responding to suggestions that LLM prompting can be a proxy for surveying human participants, previous studies on legal interpretation have argued that LLMs are unreliable for such semantic judgments due to their high degree of prompt sensitivity. However, these studies did not directly compare humans and LLMs on identical tasks, nor did they examine the performance of so-called “reasoning” models. The current study addresses these gaps in the literature by directly comparing the robustness of human and LLM judgments (with and without “reasoning”) in a legal interpretation task. Our results show that human responses were highly stable across prompt variations while LLMs were more sensitive to irrelevant prompt features. Enabling “reasoning” improved the stability of LLM responses across prompt variations. However, even “reasoning” model outputs had only moderate correlations with human judgments, and all models sometimes made decisions that no humans did in response to the same prompt. We conclude that while “reasoning” decreases prompt sensitivity, LLMs are still poor proxies for human metalinguistic judgments.

Sense and Sensitivity: “Reasoning” Models are More Robust, but can Diverge from Human Consensus in a Legal Interpretation Task

Children’s acquisition of filler-gap dependencies has been argued by some to depend on innate grammatical knowledge, while others suggest that the distributional evidence available in child-directed speech suffices. Unfortunately, the relevant input is difficult to quantify at scale with fine granularity, making this question difficult to resolve. We present a system that identifies three core filler-gap constructions in spoken English corpora -- matrix wh-questions, embedded wh-questions, and relative clauses -- and further identifies the extraction site (i.e., subject vs. object vs. adjunct). Our approach combines constituency and dependency parsing, leveraging their complementary strengths for construction classification and extraction site identification. We validate the system on human-annotated data and find that it scores well across most categories. Applying the system to 57 English CHILDES corpora, we are able to characterize children’s filler-gap input and their filler-gap production trajectories over the course of development, including construction-specific frequencies and extraction-site asymmetries. The resulting fine-grained labels enable future work in both acquisition and computational studies, which we demonstrate with a case study using filtered corpus training with language models.

Downloads

Next from ACL 2026

Uncovering Autoregressive LLM Knowledge of Thematic Fit in Event Representation

Similar lecture

Interpreting Style Representations via Style-Eliciting Prompts

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2026

Uncovering Autoregressive LLM Knowledge of Thematic Fit in Event Representation

Similar lecture

Interpreting Style Representations via Style-Eliciting Prompts

Downloads