Findings of the Second BabyLM Challenge: Sample-Efficient Pretraining on Developmentally Plausible Corpora

The BabyLM Challenge is a community effort to close the data-efficiency gap between human and computational language learners. Participants compete to optimize language model training on a fixed language data budget of 100 million words or less. This year, we released improved text corpora, as well as a vision-and-language corpus to facilitate research into cognitively plausible vision language models. Submissions were compared on evaluation tasks targeting grammatical ability, (visual) question answering, pragmatic abilities, and grounding, among other abilities. Participants could submit to a 10M-word text-only track, a 100M-word text-only track, and/or a 100M-word and image multimodal track. From 31 submissions employing diverse methods, a hybrid causal-masked language model architecture outperformed other approaches. No submissions outperformed the baselines in the multimodal track. In follow-up analyses, we found a strong relationship between training FLOPs and average performance across tasks, and that the best-performing submissions proposed changes to the training data, training objective, and model architecture. This year's BabyLM Challenge shows that there is still significant room for innovation in this setting, in particular for image-text modeling, but community-driven research can yield actionable insights about effective strategies for small-scale language modeling.

What Goes Into a LM Acceptability Judgment? Rethinking the Impact of Frequency and Length

When comparing the linguistic capabilities of language models (LMs) with humans using LM probabilities, factors such as the length of the sequence and the unigram frequency of lexical items have a significant effect on LM probabilities in ways that humans are largely robust to. Prior works in comparing LM and human acceptability judgments treat these effects uniformly across models, making a strong assumption that models require the same degree of adjustment to control for length and unigram frequency effects. We propose MORCELA, a new linking theory between LM scores and acceptability judgments where the optimal level of adjustment for these effects is estimated from data via learned parameters for length and unigram frequency. We first show that MORCELA outperforms a commonly used linking theory for acceptability--SLOR (Pauls and Klein, 2012; Lau et al. 2017)--across two families of transformer LMs (Pythia and OPT). Furthermore, we demonstrate that the assumed degrees of adjustment in SLOR for length and unigram frequency overcorrect for these confounds, and that larger models require a lower relative degree of adjustment for unigram frequency, though a significant amount of adjustment is still necessary for all models. Finally, our subsequent analysis shows that larger LMs' lower susceptibility to frequency effects can be explained by an ability to better predict rarer words in context.

How Does Code Pretraining Affect Language Model Task Performance?

Large language models are increasingly trained on corpora containing both natural language and non-linguistic data like source code. Aside from aiding programming-related tasks, anecdotal evidence suggests that including code in pretraining corpora may improve performance on other, unrelated tasks, yet to date no work has been able to establish a causal connection by controlling between language and code data. Here we do just this. We pretrain language models on datasets which interleave natural language and code in two different settings: additive, in which the total volume of data seen during pretraining is held constant; and competitive, in which the volume of language data is held constant. We study how the pretraining mixture affects performance on (a) a diverse collection of tasks included in the BigBench benchmark, and (b) compositionality, measured by generalization accuracy on semantic parsing and syntactic transformations. We find that pretraining on higher proportions of code improves performance on compositional tasks involving structured output (like semantic parsing), and mathematics. Conversely, increase code mixture can harm performance on other tasks, including on tasks that requires sensitivity to linguistic structure such as syntax or morphology, and tasks measuring real-world knowledge.

Testing learning hypotheses using neural networks by manipulating learning data

Although passivization is productive in English, it is not completely general -- some exceptions exist (e.g. *One hour was lasted by the meeting). How do English speakers learn these exceptions to an otherwise general pattern? Using neural network language models as theories of acquisition, we explore the sources of indirect evidence that a learner can leverage to learn whether a verb can passivize. We first characterize English speakers' judgments of exceptions to the passive, confirming that speakers find some verbs more passivizable than others. We then show that a neural network language model can learn restrictions to the passive that are similar to those displayed by humans, suggesting that evidence for these exceptions is available in the linguistic input. We test the causal role of two hypotheses for how the language model learns these restrictions by training models on modified training corpora, which we create by altering the existing training corpora to remove features of the input implicated by each hypothesis. We find that while the frequency with which a verb appears in the passive significantly affects its passivizability, the semantics of the verb does not. This study highlight the utility of altering a language model's training data for answering questions where complete control over a learner's input is vital.

[Call for Papers] The 2nd BabyLM Challenge: Sample-efficient pretraining on a developmentally plausible corpus

After last year's successful BabyLM Challenge, the competition will be hosted again in 2024/2025. The overarching goals of the challenge remain the same; however, some of the competition rules will be different. The big changes for this year's competition are as follows: First, we replace the loose track with a paper track, which allows (for example) non-model-based submissions, novel cognitively-inspired benchmarks, or analysis techniques. Second, we are relaxing the rules around pretraining data, and will now allow participants to construct their own datasets provided they stay within the 100M-word or 10M-word budget. Third, we introduce a multimodal vision-and-language track, and will release a corpus of 50% text-only and 50% image-text multimodal data as a starting point for LM model training. The purpose of this CfP is to provide rules for this year's challenge, explain these rule changes and their rationale in greater detail, give a timeline of this year's competition, and provide answers to frequently asked questions from last year's challenge.

SPAWNing Structural Priming Predictions from a Cognitively Motivated Parser

Structural priming is a widely used psycholinguistic paradigm to study human sentence representations. In this work we propose a framework for using empirical priming patterns to build a theory characterizing the structural representations humans construct when processing sentences. This framework uses a new cognitively motivated parser, SPAWN, to generate quantitative priming predictions from theoretical syntax and evaluate these predictions with empirical human behavior. As a case study, we apply this framework to study reduced relative clause representations in English. We use SPAWN to generate priming predictions from two theoretical accounts which make different assumptions about the structure of relative clauses. We find that the predictions from only one of these theories (Participial-Phase) align with empirical priming patterns, thus highlighting which assumptions about relative clause better capture human sentence representations.

Can You Learn Semantics Through Next-Word Prediction? The Case of Entailment

Do LMs infer the semantics of text from co-occurrence patterns in their training data? Merrill et al. (2022) argue that, in theory, probabilities predicted by an optimal LM encode semantic information about entailment relations, but it is unclear whether neural LMs trained on corpora learn entailment in this way because of strong idealizing assumptions made by Merrill et al. In this work, we investigate whether their theory can be used to decode entailment judgments from neural LMs. We find that a test similar to theirs can decode entailment relations between natural sentences, well above random chance, though not perfectly, across many datasets and LMs. This suggests LMs implicitly model aspects of semantics to predict semantic effects on sentence co-occurrence patterns. However, we find the test that predicts entailment in practice works in the opposite direction to the theoretical test. We thus revisit the assumptions underlying the original test, finding its derivation did not adequately account for redundancy in human-written text. We argue that correctly accounting for redundancy related to explanations might derive the observed flipped test and, more generally, improve linguistic theories of human speakers.

In-context Learning Generalizes, But Not Always Robustly: The Case of Syntax

In-context learning (ICL) is now a common method for supervising large language models (LLMs): given labeled examples in the input context, the LLM learns to perform the task without weight updates. Despite ICL's prevalence and utility, we understand little about whether models supervised in this manner represent the underlying structure of their tasks, rather than superficial heuristics that only generalize to identically distributed examples. In this study, we investigate the robustness of LLMs supervised via ICL using the test case of sensitivity to syntax, which is a prerequisite for robust language understanding. Our experiments are based on two simple and well-controlled syntactic transformations tasks, where correct out-of-distribution generalization requires an accurate syntactic analysis of the input. We further investigate whether out-of-distribution generalization can be improved via chain-of-thought prompting, where the model is provided with a sequence of intermediate computation steps that illustrate how the task ought to be performed. In experiments with models from the GPT, PaLM, and Llama 2 families, we find large variance across LMs on this fundamental linguistic phenomenon, and that the variance is explained more by the composition of the pre-training corpus and supervision methods than by model size. In particular, we find evidence that models pre-trained on code generalize better, and benefit to a greater extent from chain-of-thought prompting.

A Systematic Comparison of Syllogistic Reasoning in Humans and Language Models

A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans' inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate these biases, or are they able to overcome them? Focusing on the case of syllogisms -- inferences from two simple premises, which have been studied extensively in psychology -- we show that larger models are more logical than smaller ones, and also more logical than humans. At the same time, even the largest models make systematic errors, some of which mirror human reasoning biases such as ordering effects and logical fallacies. Overall, we find that language models mimic the human biases included in their training data, but are able to overcome them in some cases.

The Impact of Depth and Width on Transformer Language Model Generalization

To process novel sentences, language models (LMs) must generalize compositionally -- combine familiar elements in new ways. What aspects of a model's structure promote compositional generalization? Focusing on transformers, we test the hypothesis, motivated by recent theoretical and empirical work, that transformers generalize more compositionally when they are deeper (have more layers). Because simply adding layers increases the total number of parameters, confounding depth and size, we construct three classes of models which trade off depth for width such that the total number of parameters is kept constant (41M, 134M and 374M parameters). We pretrain all models as LMs and fine-tune them on tasks that test for compositional generalization. We report three main conclusions: (1) after fine-tuning, deeper models generalize better out-of-distribution than shallower models do, but the relative benefit of additional layers diminishes rapidly; (2) within each family, deeper models show better language modeling performance, but returns are similarly diminishing; (3) the benefits of depth for compositional generalization cannot be attributed solely to better performance on language modeling or on in-distribution data.