United States

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&#39;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&#39;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.

ACL 2026

Addressing the Ecological Fallacy in Larger LMs with Human Context

large human language models

human llms

ecological fallacy

author context

human-context-aware lms

human language models

human context

llms

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

ACL 2024

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

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.

workshop paper

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. 
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) 
<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>

To access this event page, you need to log in with the **email address you registered with**. 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.

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.

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.

What Exactly do Children Receive in Language Acquisition? A Case Study on CHILDES with Automated Detection of Filler-Gap Dependencies

Reinforcement Learning with Human Feedback (RLHF) is an increasingly popular post-training procedure for Large Language Models (LLMs) to better align outputs with human preferences. 
Therefore, one might expect some sense of human-like audience design to be induced into LLMs. 
However, RLHF and other post-training alignment methods have many complex effects on the outputs of LLMs. 
We apply an information-theoretic lens to investigate the changes in the "naturalness" of language and the presence of audience design in LLMs before and after post-training.
The Uniform Information Density (UID) Hypothesis posits that humans optimize language production and comprehension across a noisy channel by transferring information at a more uniform rate.
Accordingly, we analyze and compare how information is distributed within model-generated and human-generated text belonging to various domains to investigate the presence and form of audience design in LLMs. 
We find that pretrained and post-trained LLMs both show superhuman uniformity across various text domains, while RLHF encourages slightly more human-like, i.e., less uniform, outputs. 
However, other post-training approaches have a similar effect, suggesting that information uniformity is not a significant driver of human preferences.

An Information-Theoretic Study of RLHF-Induced Uniformity in Large Language Model Outputs

Grasping the semantics of rare constructions (form--meaning pairings) has been shown to be a challenging problem that has currently only been solved by the largest LLMs. It remains an open question if open-source models have robust constructional understanding, and if so, what learning dynamics underlie the acquisition of this knowledge. Focusing on a set of rare Paired-Focus constructions in English (e.g. "let alone", "much less"), we construct a novel dataset to test their meanings using both scalar adjectival semantics and general world knowledge. Testing a wide range of models differing in parameter count, architecture, and pretraining dataset size, we find that several modestly sized models are sensitive to both the forms and the meanings of Paired-Focus constructions. Turning to training dynamics for a set of open-checkpoint models, we find that Paired-Focus understanding emerges later in training than Paired-Focus syntactic knowledge, and that learning of Paired-Focus semantics is correlated with gains in some domains of world knowledge. Overall, our empirical results support the conclusion that modestly sized open-source models can grasp the rare Paired-Focus constructions, and demonstrate a connection between knowledge of Paired-Focus constructions and other meaning domains.

Language Models Learn Constructional Semantics, Not To Mention Syntax: Investigating LM Understanding of Paired-Focus Constructions

Visual Word Sense Disambiguation (Visual-WSD) requires ranking the correct image for an ambiguous word given a short trigger phrase. For low-resource languages, it is bottle\-necked by scarce sense-level benchmarks and limited sense-aligned multimodal supervision. We study Ukrainian and (i) extend the Ukrainian Visual-WSD benchmark from 87 to 381 instances and benchmark multilingual CLIP checkpoints and multimodal large models, and (ii) introduce two scalable Wikipedia-derived dataset construction methods. Using compute-efficient adaptation we fine-tune a multilingual CLIP backbone and show that sense-grounded supervision drives the improvements: combining our two Wikipedia-derived datasets improves HIT@1 from 37.00\% to 43.05\%.

From Sparse to Sense-Grounded: Wikipedia Training for Ukrainian Visual-WSD

Hope speech has played a vital role in online communities, yet most NLP work has focused on English and a few high-resource languages, leaving code-mixed varieties like Tulu largely unexplored. In the Shared Task on Hope Speech Detection in Code-Mixed Tulu at DravidianLangTech@ACL 2026, we have tackled two subtasks: (i) coarse-grained classification into Encouraging, Discouraging, Uninvolved and Blended categories (Task 1) and (ii) fine-grained classification into Optimistic, Realistic, Inspiring, Fading and Hopelessness (Task 2). We have fine-tuned three multilingual transformer encoders XLM-RoBERTa-base, MuRIL and mBERT on the official training splits. In Task 1, a three-way soft-voting ensemble of all three models has yielded the best performance with a macro F1 of 0.58, securing 1st place. In Task 2, XLM-RoBERTa-base alone has outperformed both MuRIL and mBERT, achieving a macro F1 of 0.42 and also securing 1st place.

CUET_SYNTHETICA@DravidianLangTech 2026: Multilingual Transformer Based Hope Speech Detection for Coarse and Fine-Grained Classification in Tulu

Abusive language targeting women has been a serious problem on Tamil social media and building systems to detect it automatically is harder than it looks. Tamil is morphologically complex, people have written it mixed with English in ways no dictionary has accounted for and a lot of the hostility has been indirect enough that has slipped past models trained on surface patterns. In the Shared Task on Abusive Tamil Text Targeting Women on Social Media DravidianLangTech@ACL 2026, we have worked on classifying Tamil YouTube comments as Abusive or Non-Abusive. We have trained three transformer models four times each with different learning rates, giving us 12 models total. Their predicted probabilities have been averaged to make the final decision. The 12-model ensemble has achieved a macro F1 of 0.8086, outperforming all individual models and securing 4th place in the shared task. Combining Tamil-specialized and multilingual transformer models has outperformed any single-architecture approach.

CUET_SYNTHETICA@DravidianLangTech 2026: Multi Architecture Transformer Ensemble for Detecting Abusive Tamil Text Targeting Women

This paper describes Team CHMOD_777's system for the DravidianLangTech@ACL 2026 shared task on political multiclass sentiment analysis of Tamil Twitter comments. The task requires classifying Tamil political tweets into seven sentiment categories under severe class imbalance (8:1 ratio). We address this challenge through LLM-based data augmentation using Gemini 2.5 Flash, expanding training data from 4,352 to 15,316 samples (3.5x the original). Our best system, MuRIL fine-tuned on augmented data with Focal Loss (gamma=3.0) and weighted sampling, achieves 35.79% Macro F1 on the development set, a 67% relative improvement over the non-augmented baseline. On the official test set, our system achieves 34.25% Macro F1, ranking 12th out of 22 participating teams. We find that (1) language-specific pre-training (MuRIL, 236M) outperforms larger general models (IndicBERT-v3, 1B), (2) smaller models benefit disproportionately from augmentation, and (3) Substantiated is the hardest category (F1=10.7%) due to its requirement for factual reasoning.

Downloads

Next from ACL 2026

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

Similar lecture

Large Language Models are Few-Shot Training Example Generators: A Case Study in Fallacy Recognition

.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

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

Similar lecture

Large Language Models are Few-Shot Training Example Generators: A Case Study in Fallacy Recognition

Downloads