Uncovering the Computational Ingredients of Human-Like Representations in LLMs

• URL: http://arxiv.org/abs/2510.01030v1
• Date: Wed, 01 Oct 2025 15:37:19 GMT
• Title: Uncovering the Computational Ingredients of Human-Like Representations in LLMs
• Authors: Zach Studdiford, Timothy T. Rogers, Kushin Mukherjee, Siddharth Suresh,
• Abstract summary: It remains unclear which of these ingredients are most crucial for building models that develop human-like representations.<n>Most current benchmarks are not suited to measuring representational alignment between humans and models.
• Score: 8.00888290370075
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The ability to translate diverse patterns of inputs into structured patterns of behavior has been thought to rest on both humans' and machines' ability to learn robust representations of relevant concepts. The rapid advancement of transformer-based large language models (LLMs) has led to a diversity of computational ingredients -- architectures, fine tuning methods, and training datasets among others -- but it remains unclear which of these ingredients are most crucial for building models that develop human-like representations. Further, most current LLM benchmarks are not suited to measuring representational alignment between humans and models, making benchmark scores unreliable for assessing if current LLMs are making progress towards becoming useful cognitive models. We address these limitations by first evaluating a set of over 70 models that widely vary in their computational ingredients on a triplet similarity task, a method well established in the cognitive sciences for measuring human conceptual representations, using concepts from the THINGS database. Comparing human and model representations, we find that models that undergo instruction-finetuning and which have larger dimensionality of attention heads are among the most human aligned, while multimodal pretraining and parameter size have limited bearing on alignment. Correlations between alignment scores and scores on existing benchmarks reveal that while some benchmarks (e.g., MMLU) are better suited than others (e.g., MUSR) for capturing representational alignment, no existing benchmark is capable of fully accounting for the variance of alignment scores, demonstrating their insufficiency in capturing human-AI alignment. Taken together, our findings help highlight the computational ingredients most essential for advancing LLMs towards models of human conceptual representation and address a key benchmarking gap in LLM evaluation.

Related papers

• HumanLLM: Towards Personalized Understanding and Simulation of Human Nature [72.55730315685837]
  HumanLLM is a foundation model designed for personalized understanding and simulation of individuals.<n>We first construct the Cognitive Genome, a large-scale corpus curated from real-world user data on platforms like Reddit, Twitter, Blogger, and Amazon.<n>We then formulate diverse learning tasks and perform supervised fine-tuning to empower the model to predict a wide range of individualized human behaviors, thoughts, and experiences.
  arXiv  Detail & Related papers  (2026-01-22T09:27:27Z)
• Beyond Accuracy: Characterizing Code Comprehension Capabilities in (Large) Language Models [4.841487377596519]
  This paper investigates whether Large Language Models' code-comprehension performance aligns with traditional human-centric software metrics.<n>We introduce a diagnostic framework that reframes code understanding as a binary input-output consistency task.
  arXiv  Detail & Related papers  (2026-01-19T10:58:24Z)
• Computational Turing Test Reveals Systematic Differences Between Human and AI Language [0.0]
  Large language models (LLMs) are increasingly used in the social sciences to simulate human behavior.<n>Existing validation efforts rely heavily on human-judgment-based evaluations.<n>This paper introduces a computational Turing test to assess how closely LLMs approximate human language.
  arXiv  Detail & Related papers  (2025-11-06T08:56:37Z)
• From Tokens to Thoughts: How LLMs and Humans Trade Compression for Meaning [63.25540801694765]
  Large Language Models (LLMs) demonstrate striking linguistic abilities, yet whether they achieve this same balance remains unclear.<n>We apply the Information Bottleneck principle to quantitatively compare how LLMs and humans navigate this compression-meaning trade-off.
  arXiv  Detail & Related papers  (2025-05-21T16:29:00Z)
• Model Utility Law: Evaluating LLMs beyond Performance through Mechanism Interpretable Metric [99.56567010306807]
  Large Language Models (LLMs) have become indispensable across academia, industry, and daily applications.<n>One core challenge of evaluation in the large language model (LLM) era is the generalization issue.<n>We propose Model Utilization Index (MUI), a mechanism interpretability enhanced metric that complements traditional performance scores.
  arXiv  Detail & Related papers  (2025-04-10T04:09:47Z)
• A Statistical Framework for Ranking LLM-Based Chatbots [57.59268154690763]
  We propose a statistical framework that incorporates key advancements to address specific challenges in pairwise comparison analysis.<n>First, we introduce a factored tie model that enhances the ability to handle groupings of human-judged comparisons.<n>Second, we extend the framework to model covariance tiers between competitors, enabling deeper insights into performance relationships.<n>Third, we resolve optimization challenges arising from parameter non-uniqueness by introducing novel constraints.
  arXiv  Detail & Related papers  (2024-12-24T12:54:19Z)
• A Flexible Method for Behaviorally Measuring Alignment Between Human and Artificial Intelligence Using Representational Similarity Analysis [0.1957338076370071]
  We adapted Representational Similarity Analysis (RSA), a method that uses pairwise similarity ratings to quantify alignment between AIs and humans.<n>We tested this approach on semantic alignment across text and image modalities, measuring how different Large Language and Vision Language Model (LLM and VLM) similarity judgments aligned with human responses at both group and individual levels.
  arXiv  Detail & Related papers  (2024-11-30T20:24:52Z)
• SymbolicAI: A framework for logic-based approaches combining generative models and solvers [9.841285581456722]
  We introduce SymbolicAI, a versatile and modular framework employing a logic-based approach to concept learning and flow management in generative processes. We treat large language models (LLMs) as semantic solvers that execute tasks based on both natural and formal language instructions.
  arXiv  Detail & Related papers  (2024-02-01T18:50:50Z)
• CLOMO: Counterfactual Logical Modification with Large Language Models [109.60793869938534]
  We introduce a novel task, Counterfactual Logical Modification (CLOMO), and a high-quality human-annotated benchmark. In this task, LLMs must adeptly alter a given argumentative text to uphold a predetermined logical relationship. We propose an innovative evaluation metric, the Self-Evaluation Score (SES), to directly evaluate the natural language output of LLMs.
  arXiv  Detail & Related papers  (2023-11-29T08:29:54Z)
• Evaluating Representations with Readout Model Switching [19.907607374144167]
  In this paper, we propose to use the Minimum Description Length (MDL) principle to devise an evaluation metric. We design a hybrid discrete and continuous-valued model space for the readout models and employ a switching strategy to combine their predictions. The proposed metric can be efficiently computed with an online method and we present results for pre-trained vision encoders of various architectures.
  arXiv  Detail & Related papers  (2023-02-19T14:08:01Z)
• Conceptually Diverse Base Model Selection for Meta-Learners in Concept Drifting Data Streams [3.0938904602244355]
  We present a novel approach for estimating the conceptual similarity of base models, which is calculated using the Principal Angles (PAs) between their underlying subspaces. We evaluate these methods against thresholding using common ensemble pruning metrics, namely predictive performance and Mutual Information (MI) in the context of online Transfer Learning (TL) Our results show that conceptual similarity thresholding has a reduced computational overhead, and yet yields comparable predictive performance to thresholding using predictive performance and MI.
  arXiv  Detail & Related papers  (2021-11-29T13:18:53Z)
• Prototypical Contrastive Learning of Unsupervised Representations [171.3046900127166]
  Prototypical Contrastive Learning (PCL) is an unsupervised representation learning method. PCL implicitly encodes semantic structures of the data into the learned embedding space. PCL outperforms state-of-the-art instance-wise contrastive learning methods on multiple benchmarks.
  arXiv  Detail & Related papers  (2020-05-11T09:53:36Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.