Towards Improved Sentence Representations using Token Graphs

• URL: http://arxiv.org/abs/2603.03389v1
• Date: Tue, 03 Mar 2026 09:00:01 GMT
• Title: Towards Improved Sentence Representations using Token Graphs
• Authors: Krishna Sri Ipsit Mantri, Carola-Bibiane Schönlieb, Zorah Lähner, Moshe Eliasof,
• Abstract summary: We introduce GLOT, a structure-aware pooling module that reframes pooling as relational learning followed by aggregation.<n>On a diagnostic stress test where 90% of tokens are random distractors, GLOT maintains over 97% accuracy while baseline methods collapse.<n>It is competitive with state-of-the-art techniques on benchmarks like GLUE and MTEB with 20x fewer trainable parameters and speeds up the training time by over 100x compared with parameter-efficient fine-tuning methods.
• Score: 41.412173502714225
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Obtaining a single-vector representation from a Large Language Model's (LLM) token-level outputs is a critical step for nearly all sentence-level tasks. However, standard pooling methods like mean or max aggregation treat tokens as an independent set, discarding the rich relational structure captured by the model's self-attention layers and making them susceptible to signal dilution. To address this, we introduce GLOT, a lightweight, structure-aware pooling module that reframes pooling as relational learning followed by aggregation. Operating on the outputs of a frozen LLM, GLOT first constructs a latent token-similarity graph, then refines token representations with a graph neural network, and finally aggregates them using a readout layer. Experimentally, our approach is remarkably robust and efficient: on a diagnostic stress test where 90% of tokens are random distractors, GLOT maintains over 97% accuracy while baseline methods collapse. Furthermore, it is competitive with state-of-the-art techniques on benchmarks like GLUE and MTEB with 20x fewer trainable parameters and speeds up the training time by over 100x compared with parameter-efficient fine-tuning methods. Supported by a theoretical analysis of its expressive power, our work shows that learning over token graphs is a powerful paradigm for the efficient adaptation of frozen LLMs. Our code is published at https://github.com/ipsitmantri/GLOT.

Related papers

• Semi-supervised Instruction Tuning for Large Language Models on Text-Attributed Graphs [62.544129365882014]
  We propose a novel Semi-supervised Instruction Tuning pipeline for Graph Learning, named SIT-Graph.<n> SIT-Graph is model-agnostic and can be seamlessly integrated into any graph instruction tuning method that utilizes LLMs as the predictor.<n>Extensive experiments demonstrate that when incorporated into state-of-the-art graph instruction tuning methods, SIT-Graph significantly enhances their performance on text-attributed graph benchmarks.
  arXiv  Detail & Related papers  (2026-01-19T08:10:53Z)
• Improving LLM Reasoning with Homophily-aware Structural and Semantic Text-Attributed Graph Compression [55.51959317490934]
  Large language models (LLMs) have demonstrated promising capabilities in Text-Attributed Graph (TAG) understanding.<n>We argue that graphs inherently contain rich structural and semantic information, and that their effective exploitation can unlock potential gains in LLMs reasoning performance.<n>We propose Homophily-aware Structural and Semantic Compression for LLMs (HS2C), a framework centered on exploiting graph homophily.
  arXiv  Detail & Related papers  (2026-01-13T03:35:18Z)
• GILT: An LLM-Free, Tuning-Free Graph Foundational Model for In-Context Learning [50.40400074353263]
  Graph Neural Networks (GNNs) are powerful tools for precessing relational data but often struggle to generalize to unseen graphs.<n>We introduce textbfGraph textbfIn-context textbfL textbfTransformer (GILT), a framework built on an LLM-free and tuning-free architecture.
  arXiv  Detail & Related papers  (2025-10-06T08:09:15Z)
• GraphLAMA: Enabling Efficient Adaptation of Graph Language Models with Limited Annotations [46.15515676751084]
  Large language models (LLMs) have demonstrated their strong capabilities in various domains, and have been recently integrated for graph analysis as graph language models (GLMs)<n>With LLMs as the predictor, some GLMs can interpret unseen tasks described by natural language, and learn from a few examples in the prompts without parameter tuning, known as in-context learning (ICL)<n>We propose GraphLAMA method, with its model backbone and learning schemes specialized for efficient tuning and inference.
  arXiv  Detail & Related papers  (2025-06-11T16:38:01Z)
• Training-Free Tokenizer Transplantation via Orthogonal Matching Pursuit [45.18582668677648]
  We present a training-free method to transplant tokenizers in large language models.<n>We approximate each out-of-vocabulary token as a sparse linear combination of shared tokens.<n>We show that OMP achieves best zero-shot preservation of the base model's performance.
  arXiv  Detail & Related papers  (2025-06-07T00:51:27Z)
• Scalability Matters: Overcoming Challenges in InstructGLM with Similarity-Degree-Based Sampling [1.2805157669888096]
  We propose SDM-InstructGLM, a novel instruction-tuned Graph Language Model (InstructGLM) framework that enhances scalability and efficiency without relying on GNNs.<n>Our method introduces a similarity-degree-based biased random walk mechanism, which selectively samples and encodes graph information based on node-feature similarity and degree centrality.<n>Our results demonstrate the feasibility of LLM-only graph processing, enabling scalable and interpretable Graph Language Models (GLMs) optimized through instruction-based fine-tuning.
  arXiv  Detail & Related papers  (2025-05-02T06:08:21Z)
• Compile Scene Graphs with Reinforcement Learning [69.36723767339001]
  Next-token prediction is the fundamental principle for training large language models (LLMs)<n>We introduce R1-SGG, a multimodal LLM (M-LLM) initially trained via supervised fine-tuning (SFT) on the scene graph dataset.<n>We design a set of graph-centric rewards, including three recall-based variants -- Hard Recall, Hard Recall+Relax, and Soft Recall.
  arXiv  Detail & Related papers  (2025-04-18T10:46:22Z)
• Exact Byte-Level Probabilities from Tokenized Language Models for FIM-Tasks and Model Ensembles [23.134664392314264]
  Tokenization is associated with many poorly understood shortcomings in language models (LMs)<n>This work studies how tokenization impacts model performance by analyzing and comparing models with their byte-level counterparts.<n>We introduce the Byte-Token Representation Lemma, a framework that establishes a mapping between the learned token distribution and its equivalent byte-level distribution.
  arXiv  Detail & Related papers  (2024-10-11T23:30:42Z)
• GLBench: A Comprehensive Benchmark for Graph with Large Language Models [41.89444363336435]
  We introduce GLBench, the first comprehensive benchmark for evaluating GraphLLM methods in both supervised and zero-shot scenarios. GLBench provides a fair and thorough evaluation of different categories of GraphLLM methods, along with traditional baselines such as graph neural networks.
  arXiv  Detail & Related papers  (2024-07-10T08:20:47Z)
• Combining Label Propagation and Simple Models Out-performs Graph Neural Networks [52.121819834353865]
  We show that for many standard transductive node classification benchmarks, we can exceed or match the performance of state-of-the-art GNNs. We call this overall procedure Correct and Smooth (C&S) Our approach exceeds or nearly matches the performance of state-of-the-art GNNs on a wide variety of benchmarks.
  arXiv  Detail & Related papers  (2020-10-27T02:10:52Z)

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.