Where Does Social Reasoning Come From? Capability Provenance in Language Models

• URL: http://arxiv.org/abs/2606.19625v1
• Date: Wed, 17 Jun 2026 22:06:19 GMT
• Title: Where Does Social Reasoning Come From? Capability Provenance in Language Models
• Abstract summary: We map which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B.<n>Training-data attribution measures how strongly each training document influences a model's predictions on a benchmark.<n>We compute gradient-based attribution over a working set drawn from the de-duplicated Dolma3 mix.
• Score: 11.7652444083388
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B. Training-data attribution measures how strongly each training document influences a model's predictions on a benchmark, but document-level scores are too noisy to identify which corpus regions support which capabilities, and prior work has emphasized factual knowledge rather than reasoning. We compute gradient-based attribution (TrackStar via Bergson) over a working set drawn from the de-duplicated Dolma3 mix, aggregate influence across WebOrganizer's 24-format x 24-topic taxonomy (576 bins), and contrast benchmark pairs in a 2x2 design that varies domain (social vs. STEM) and capability type (reasoning vs. knowledge): SocialIQA and MMLU Social Sciences against ARC-Challenge and MMLU STEM. Social and STEM reasoning draw on qualitatively distinct corpus regions, and the contrast is sharper at the reasoning level than at the knowledge level. Targeted machine unlearning provides partial causal validation: forgetting high-attribution topic bins (e.g., Literature for SocialIQA) degrades the aligned benchmark more than within-bin random baselines, and we open-source all code, sampling manifests, the bin-level influence matrix, and unlearning checkpoints.

Related papers

• Perceive, Refine, Reason: A Calibrated Pipeline for Measuring Indicators in Strategic Visual Communication on Social Media [2.927717169939877]
  We introduce Perceive, Refine, Reason (PRR), a calibrated pipeline that turns flexible vision-language detectors into auditable measurement instruments for social-scientific research.<n>PRR combines natural-language category prompts with pixel-level spatial refinement via the Segment Anything Model (SAM)<n>A complementary three-tier auditability framework applies learning to profile per-category reliability, support task-aligned configuration, and statistically correct prevalence estimates.
  arXiv  Detail & Related papers  (2026-09-13T17:54:01Z)
• Does Syntax Matter? A Graph-Augmented Variational Topic Model for Computational Social Sciences [51.56484100374058]
  This paper introduces the Structural Contextual Probabilistic Topic Model ( SCPTM), an architecture that incorporates syntactic dependency relations into topic inference.<n>We evaluate seven topic modeling techniques across four corpora differing in register and discourse structure.<n>Results show that SCPTM's neural architecture yields substantial gains in document-topic alignment over generative baselines.
  arXiv  Detail & Related papers  (2026-09-07T17:39:06Z)
• Unveiling Public Opinion: A Study of Sentiment Analysis Using LSTM and Traditional Models [0.3655021726150367]
  Sentiment analysis on Twitter employs sophisticated computational techniques to categorize tweets into positive, negative, or neutral sentiments.<n>This paper examines the effectiveness of various machine learning and deep learning approaches.<n>It identifies the optimal sentiment analysis model using a Kaggle Twitter dataset that has been preprocessed through tokenization, lemmatization, and stopword elimination.
  arXiv  Detail & Related papers  (2026-07-08T16:32:39Z)
• A Survey of Machine Learning Models and Datasets for the Multi-label Classification of Textual Hate Speech in English [3.216132991084434]
  This work presents the first systematic and comprehensive survey of scientific literature on this emerging research landscape in English.<n>We contribute with a concise overview of 28 datasets suited for training multi-label classification models.<n>We identify imbalanced training data, reliance on crowdsourcing platforms, small and sparse datasets, and missing methodological alignment.
  arXiv  Detail & Related papers  (2025-04-11T15:16:31Z)
• A Top-down Graph-based Tool for Modeling Classical Semantic Maps: A Crosslinguistic Case Study of Supplementary Adverbs [50.982315553104975]
  Semantic map models (SMMs) construct a network-like conceptual space from cross-linguistic instances or forms.<n>Most SMMs are manually built by human experts using bottom-up procedures.<n>We propose a novel graph-based algorithm that automatically generates conceptual spaces and SMMs in a top-down manner.
  arXiv  Detail & Related papers  (2024-12-02T12:06:41Z)
• How to Handle Different Types of Out-of-Distribution Scenarios in Computational Argumentation? A Comprehensive and Fine-Grained Field Study [59.13867562744973]
  This work systematically assesses LMs' capabilities for out-of-distribution (OOD) scenarios. We find that the efficacy of such learning paradigms varies with the type of OOD. Specifically, while ICL excels for domain shifts, prompt-based fine-tuning surpasses for topic shifts.
  arXiv  Detail & Related papers  (2023-09-15T11:15:47Z)
• Bias and Fairness in Large Language Models: A Survey [73.87651986156006]
  We present a comprehensive survey of bias evaluation and mitigation techniques for large language models (LLMs) We first consolidate, formalize, and expand notions of social bias and fairness in natural language processing. We then unify the literature by proposing three intuitive, two for bias evaluation, and one for mitigation.
  arXiv  Detail & Related papers  (2023-09-02T00:32:55Z)
• Unsupervised Sentiment Analysis of Plastic Surgery Social Media Posts [91.3755431537592]
  The massive collection of user posts across social media platforms is primarily untapped for artificial intelligence (AI) use cases. Natural language processing (NLP) is a subfield of AI that leverages bodies of documents, known as corpora, to train computers in human-like language understanding. This study demonstrates that the applied results of unsupervised analysis allow a computer to predict either negative, positive, or neutral user sentiment towards plastic surgery.
  arXiv  Detail & Related papers  (2023-07-05T20:16:20Z)
• Fairness meets Cross-Domain Learning: a new perspective on Models and Metrics [80.07271410743806]
  We study the relationship between cross-domain learning (CD) and model fairness. We introduce a benchmark on face and medical images spanning several demographic groups as well as classification and localization tasks. Our study covers 14 CD approaches alongside three state-of-the-art fairness algorithms and shows how the former can outperform the latter.
  arXiv  Detail & Related papers  (2023-03-25T09:34:05Z)
• Multi-Aspect Sentiment Analysis with Latent Sentiment-Aspect Attribution [7.289918297809611]
  We introduce a new framework called the sentiment-aspect attribution module (SAAM) The framework works by exploiting the correlations between sentence-level embedding features and variations of document-level aspect rating scores. Experiments on a hotel review dataset and a beer review dataset have shown SAAM can improve sentiment analysis performance.
  arXiv  Detail & Related papers  (2020-12-15T16:34:36Z)
• Weakly-Supervised Aspect-Based Sentiment Analysis via Joint Aspect-Sentiment Topic Embedding [71.2260967797055]
  We propose a weakly-supervised approach for aspect-based sentiment analysis. We learn sentiment, aspect> joint topic embeddings in the word embedding space. We then use neural models to generalize the word-level discriminative information.
  arXiv  Detail & Related papers  (2020-10-13T21:33:24Z)

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.