Understanding Contextual Recall in Transformers: How Finetuning Enables In-Context Reasoning over Pretraining Knowledge

• URL: http://arxiv.org/abs/2603.20969v1
• Date: Sat, 21 Mar 2026 22:46:55 GMT
• Title: Understanding Contextual Recall in Transformers: How Finetuning Enables In-Context Reasoning over Pretraining Knowledge
• Abstract summary: We investigate whether contextual recall emerges from pretraining alone.<n>We show that finetuning on tasks requiring implicit inference, distinct from the ICL evaluation, triggers the emergence of contextual recall.<n>For mechanistic insight, we derive a construction for an attention-only transformer that replicates the transition from factual to contextual recall.
• Score: 50.009682083079205
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Transformer-based language models excel at in-context learning (ICL), where they can adapt to new tasks based on contextual examples, without parameter updates. In a specific form of ICL, which we refer to as \textit{contextual recall}, models pretrained on open-ended text leverage pairwise examples to recall specific facts in novel prompt formats. We investigate whether contextual recall emerges from pretraining alone, what finetuning is required, and what mechanisms drive the necessary representations. For this, we introduce a controlled synthetic framework where pretraining sequences consist of subject-grammar-attribute tuples, with attribute types tied to grammar statistics. We demonstrate that while such pretraining successfully yields factual knowledge, it is insufficient for contextual recall: models fail to implicitly infer attribute types when the grammar statistics are removed in ICL prompts. However, we show that finetuning on tasks requiring implicit inference, distinct from the ICL evaluation, using a subset of subjects, triggers the emergence of contextual recall across all subjects. This transition is accompanied by the formation of low-dimensional latent encodings of the shared attribute type. For mechanistic insight, we derive a construction for an attention-only transformer that replicates the transition from factual to contextual recall, corroborated by empirical validation.

Related papers

• Weave of Formal Thought [51.56484100374058]
  We introduce Weave of Formal Thought (WoFT), a paradigm uniting rigorous syntactic validation with learned structural representations.<n>We present a latent-variable fine-tuning method training the language model to interleave non-terminal grammar symbols directly into generation.<n>For Python, fine-tuning StarCoder2-3B with our RWS objective reduces per-token cross-entropy by 14.3% relative to a text-only SFT baseline.
  arXiv  Detail & Related papers  (2026-06-24T15:58:11Z)
• Fine-Tuning Dynamics of In-Context Factual Recall in Transformers [63.1821972855273]
  In-context learning -- performing tasks based on examples given in the prompt -- has emerged in large language models.<n>We study how incontext learning leverages factual knowledge recall.<n>We analyze the supervised fine-tuning dynamics of a one-layer transformer on IC-recall data.
  arXiv  Detail & Related papers  (2026-05-26T23:47:26Z)
• Trace Is In Sentences: Unbiased Lightweight ChatGPT-Generated Text Detector [2.11622808613962]
  We introduce a novel task to detect both original and PSP-modified AI-generated texts.<n>We propose a lightweight framework that classifies texts based on their internal structure.<n>Our approach encodes sentence embeddings from pre-trained language models and models their relationships via attention.
  arXiv  Detail & Related papers  (2025-09-23T02:00:35Z)
• SCOPE: A Self-supervised Framework for Improving Faithfulness in Conditional Text Generation [55.61004653386632]
  Large Language Models (LLMs) often produce hallucinations, i.e., information that is unfaithful or not grounded in the input context.<n>This paper introduces a novel self-supervised method for generating a training set of unfaithful samples.<n>We then refine the model using a training process that encourages the generation of grounded outputs over unfaithful ones.
  arXiv  Detail & Related papers  (2025-02-19T12:31:58Z)
• On the Loss of Context-awareness in General Instruction Fine-tuning [101.03941308894191]
  We investigate the loss of context awareness after supervised fine-tuning.<n>We find that the performance decline is associated with a bias toward different roles learned during conversational instruction fine-tuning.<n>We propose a metric to identify context-dependent examples from general instruction fine-tuning datasets.
  arXiv  Detail & Related papers  (2024-11-05T00:16:01Z)
• Context-aware Prompt Tuning: Advancing In-Context Learning with Adversarial Methods [69.36397993451742]
  This work introduces Context-aware Prompt Tuning (CPT), a method inspired by ICL, PT, and adversarial attacks. We modify specific context tokens, considering the unique structure of input and output formats. Inspired by adversarial attacks, we adjust the input based on the labels present in the context, focusing on minimizing, rather than maximizing, the loss.
  arXiv  Detail & Related papers  (2024-10-22T17:45:47Z)
• From Unstructured Data to In-Context Learning: Exploring What Tasks Can Be Learned and When [19.841163050181194]
  Large language models (LLMs) like transformers demonstrate impressive in-context learning (ICL) capabilities. We examine what enables ICL in models trained on unstructured data, focusing on critical sequence model requirements and training data structure. We find that many ICL capabilities can emerge simply from co-occurrence of semantically related word pairs in unstructured data. We identify two cases where ICL fails: one in logic reasoning tasks that require generalizing to new, unseen patterns, and another in analogy completion where relevant word pairs appear only in fixed training positions.
  arXiv  Detail & Related papers  (2024-05-31T18:46:06Z)
• Mixed-effects transformers for hierarchical adaptation [1.9105318290910576]
  We introduce the mixed-effects transformer (MET), a novel approach for learning hierarchically-structured prefixes. We show how the popular class of mixed-effects models may be extended to transformer-based architectures.
  arXiv  Detail & Related papers  (2022-05-03T19:34:15Z)
• An Explanation of In-context Learning as Implicit Bayesian Inference [117.19809377740188]
  We study the role of the pretraining distribution on the emergence of in-context learning. We prove that in-context learning occurs implicitly via Bayesian inference of the latent concept. We empirically find that scaling model size improves in-context accuracy even when the pretraining loss is the same.
  arXiv  Detail & Related papers  (2021-11-03T09:12:33Z)

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.