Comparing Bottom-Up and Top-Down Steering Approaches on In-Context Learning Tasks

• URL: http://arxiv.org/abs/2411.07213v1
• Date: Mon, 11 Nov 2024 18:36:17 GMT
• Title: Comparing Bottom-Up and Top-Down Steering Approaches on In-Context Learning Tasks
• Abstract summary: Key objective of interpretability research on large language models is to develop methods for robustly steering models toward desired behaviors. We present a case study comparing the effectiveness of representative vector steering methods from each branch. We find that ICVs outperform FVs in behavioral shifting, whereas FVs excel in tasks requiring more precision.
• Score: 7.363811256593798
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: A key objective of interpretability research on large language models (LLMs) is to develop methods for robustly steering models toward desired behaviors. To this end, two distinct approaches to interpretability -- ``bottom-up" and ``top-down" -- have been presented, but there has been little quantitative comparison between them. We present a case study comparing the effectiveness of representative vector steering methods from each branch: function vectors (FV; arXiv:2310.15213), as a bottom-up method, and in-context vectors (ICV; arXiv:2311.06668) as a top-down method. While both aim to capture compact representations of broad in-context learning tasks, we find they are effective only on specific types of tasks: ICVs outperform FVs in behavioral shifting, whereas FVs excel in tasks requiring more precision. We discuss the implications for future evaluations of steering methods and for further research into top-down and bottom-up steering given these findings.

Related papers

• PEARL: Performance-Enhanced Aggregated Representation Learning [9.98626433752485]
  This paper proposes a performance-enhanced aggregated representation learning method.<n>It combines multiple representation learning approaches to improve the performance of downstream tasks.<n>We evaluate our method on diverse tasks by comparing it with advanced machine learning models.
  arXiv  Detail & Related papers  (2025-09-29T05:50:29Z)
• USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation [24.90512145836643]
  We introduce a Unified Skeleton-based Dense Representation Learning framework based on feature decorrelation. We show that our approach significantly outperforms the current state-of-the-art (SOTA) approaches.
  arXiv  Detail & Related papers  (2024-12-12T12:20:27Z)
• Revisiting Weight Averaging for Model Merging [16.503826062785773]
  Model merging aims to build a multi-task learner by combining the parameters of individually fine-tuned models without additional training. Weight averaging implicitly induces task vectors centered around the weight averaging itself. Applying a low-rank approximation to these centered task vectors significantly improves merging performance.
  arXiv  Detail & Related papers  (2024-12-11T06:29:20Z)
• Improving General Text Embedding Model: Tackling Task Conflict and Data Imbalance through Model Merging [33.23758947497205]
  Advanced embedding models are typically developed using large-scale multi-task data and joint training across multiple tasks. To overcome these challenges, we explore model merging-a technique that combines independently trained models to mitigate gradient conflicts and balance data distribution. We introduce a novel method, Self Positioning, which efficiently searches for optimal model combinations within the space of task vectors using gradient descent.
  arXiv  Detail & Related papers  (2024-10-19T08:39:21Z)
• FlowRetrieval: Flow-Guided Data Retrieval for Few-Shot Imitation Learning [28.523528119584526]
  Few-shot imitation learning relies on only a small amount of task-specific demonstrations to efficiently adapt a policy for a given downstream tasks. We propose FlowRetrieval, an approach that leverages optical flow representations for both extracting similar motions to target tasks from prior data. Our results show FlowRetrieval significantly outperforms prior methods across simulated and real-world domains.
  arXiv  Detail & Related papers  (2024-08-29T23:48:08Z)
• Towards Understanding Sensitive and Decisive Patterns in Explainable AI: A Case Study of Model Interpretation in Geometric Deep Learning [18.408342615833185]
  This research focuses on distinguishing between two critical data patterns -- sensitive patterns (model-related) and decisive patterns (task-related) We compare the effectiveness of two main streams of interpretation methods: post-hoc methods and self-interpretable methods. Our findings indicate that post-hoc methods tend to provide interpretations better aligned with sensitive patterns, whereas certain self-interpretable methods exhibit strong and stable performance in detecting decisive patterns.
  arXiv  Detail & Related papers  (2024-06-30T22:59:15Z)
• KOPPA: Improving Prompt-based Continual Learning with Key-Query Orthogonal Projection and Prototype-based One-Versus-All [24.50129285997307]
  We introduce a novel key-query learning strategy to enhance prompt matching efficiency and address the challenge of shifting features. Our method empowers the model to achieve results surpassing those of current state-of-the-art approaches by a large margin of up to 20%.
  arXiv  Detail & Related papers  (2023-11-26T20:35:19Z)
• Synergies between Disentanglement and Sparsity: Generalization and Identifiability in Multi-Task Learning [79.83792914684985]
  We prove a new identifiability result that provides conditions under which maximally sparse base-predictors yield disentangled representations. Motivated by this theoretical result, we propose a practical approach to learn disentangled representations based on a sparsity-promoting bi-level optimization problem.
  arXiv  Detail & Related papers  (2022-11-26T21:02:09Z)
• ACP++: Action Co-occurrence Priors for Human-Object Interaction Detection [102.9428507180728]
  A common problem in the task of human-object interaction (HOI) detection is that numerous HOI classes have only a small number of labeled examples. We observe that there exist natural correlations and anti-correlations among human-object interactions. We present techniques to learn these priors and leverage them for more effective training, especially on rare classes.
  arXiv  Detail & Related papers  (2021-09-09T06:02:50Z)
• Few-Shot Fine-Grained Action Recognition via Bidirectional Attention and Contrastive Meta-Learning [51.03781020616402]
  Fine-grained action recognition is attracting increasing attention due to the emerging demand of specific action understanding in real-world applications. We propose a few-shot fine-grained action recognition problem, aiming to recognize novel fine-grained actions with only few samples given for each class. Although progress has been made in coarse-grained actions, existing few-shot recognition methods encounter two issues handling fine-grained actions.
  arXiv  Detail & Related papers  (2021-08-15T02:21:01Z)
• MCDAL: Maximum Classifier Discrepancy for Active Learning [74.73133545019877]
  Recent state-of-the-art active learning methods have mostly leveraged Generative Adversarial Networks (GAN) for sample acquisition. We propose in this paper a novel active learning framework that we call Maximum Discrepancy for Active Learning (MCDAL) In particular, we utilize two auxiliary classification layers that learn tighter decision boundaries by maximizing the discrepancies among them.
  arXiv  Detail & Related papers  (2021-07-23T06:57:08Z)
• SIMPLE: SIngle-network with Mimicking and Point Learning for Bottom-up Human Pose Estimation [81.03485688525133]
  We propose a novel multi-person pose estimation framework, SIngle-network with Mimicking and Point Learning for Bottom-up Human Pose Estimation (SIMPLE) Specifically, in the training process, we enable SIMPLE to mimic the pose knowledge from the high-performance top-down pipeline. Besides, SIMPLE formulates human detection and pose estimation as a unified point learning framework to complement each other in single-network.
  arXiv  Detail & Related papers  (2021-04-06T13:12:51Z)
• 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.