Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models

• URL: http://arxiv.org/abs/2408.00113v2
• Date: Wed, 30 Oct 2024 14:21:59 GMT
• Title: Measuring Progress in Dictionary Learning for Language Model Interpretability with Board Game Models
• Authors: Adam Karvonen, Benjamin Wright, Can Rager, Rico Angell, Jannik Brinkmann, Logan Smith, Claudio Mayrink Verdun, David Bau, Samuel Marks,
• Abstract summary: We propose to measure progress in interpretable dictionary learning by working in the setting of LMs trained on chess and Othello transcripts. We introduce a new SAE training technique, $textitp-annealing$, which improves performance on prior unsupervised metrics as well as our new metrics.
• Score: 18.77400885091398
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: What latent features are encoded in language model (LM) representations? Recent work on training sparse autoencoders (SAEs) to disentangle interpretable features in LM representations has shown significant promise. However, evaluating the quality of these SAEs is difficult because we lack a ground-truth collection of interpretable features that we expect good SAEs to recover. We thus propose to measure progress in interpretable dictionary learning by working in the setting of LMs trained on chess and Othello transcripts. These settings carry natural collections of interpretable features -- for example, "there is a knight on F3" -- which we leverage into $\textit{supervised}$ metrics for SAE quality. To guide progress in interpretable dictionary learning, we introduce a new SAE training technique, $\textit{p-annealing}$, which improves performance on prior unsupervised metrics as well as our new metrics.

Related papers

• Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders [8.003244901104111]
  We propose a regularization technique for improving feature learning by encouraging SAEs trained in parallel to learn similar features. textscMFR can improve the reconstruction loss of SAEs by up to 21.21% on GPT-2 Small, and 6.67% on EEG data.
  arXiv  Detail & Related papers  (2024-11-02T11:42:23Z)
• Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoders [115.34050914216665]
  Sparse Autoencoders (SAEs) have emerged as a powerful unsupervised method for extracting sparse representations from language models. We introduce a suite of 256 SAEs, trained on each layer and sublayer of the Llama-3.1-8B-Base model, with 32K and 128K features. We assess the generalizability of SAEs trained on base models to longer contexts and fine-tuned models.
  arXiv  Detail & Related papers  (2024-10-27T17:33:49Z)
• A Tale of Two Languages: Large-Vocabulary Continuous Sign Language Recognition from Spoken Language Supervision [74.972172804514]
  We introduce a multi-task Transformer model, CSLR2, that is able to ingest a signing sequence and output in a joint embedding space between signed language and spoken language text. New dataset annotations provide continuous sign-level annotations for six hours of test videos, and will be made publicly available. Our model significantly outperforms the previous state of the art on both tasks.
  arXiv  Detail & Related papers  (2024-05-16T17:19:06Z)
• Towards Principled Evaluations of Sparse Autoencoders for Interpretability and Control [43.860799289234755]
  We propose a framework for evaluating feature dictionaries in the context of specific tasks, by comparing them against emphmagnitude feature dictionaries. First, we demonstrate that supervised dictionaries achieve excellent approximation, control, and interpretability of model computations on the task. We apply this framework to the indirect object identification (IOI) task using GPT-2 Small, with sparse autoencoders (SAEs) trained on either the IOI or OpenWebText datasets.
  arXiv  Detail & Related papers  (2024-05-14T07:07:13Z)
• CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation [94.59630161324013]
  We propose CoAnnotating, a novel paradigm for Human-LLM co-annotation of unstructured texts at scale. Our empirical study shows CoAnnotating to be an effective means to allocate work from results on different datasets, with up to 21% performance improvement over random baseline.
  arXiv  Detail & Related papers  (2023-10-24T08:56:49Z)
• Alleviating Over-smoothing for Unsupervised Sentence Representation [96.19497378628594]
  We present a Simple method named Self-Contrastive Learning (SSCL) to alleviate this issue. Our proposed method is quite simple and can be easily extended to various state-of-the-art models for performance boosting.
  arXiv  Detail & Related papers  (2023-05-09T11:00:02Z)
• ParroT: Translating during Chat using Large Language Models tuned with Human Translation and Feedback [90.20262941911027]
  ParroT is a framework to enhance and regulate the translation abilities during chat. Specifically, ParroT reformulates translation data into the instruction-following style. We propose three instruction types for finetuning ParroT models, including translation instruction, contrastive instruction, and error-guided instruction.
  arXiv  Detail & Related papers  (2023-04-05T13:12:00Z)
• Memory Augmented Lookup Dictionary based Language Modeling for Automatic Speech Recognition [20.926163659469587]
  We propose a new memory augmented lookup dictionary based Transformer architecture for LM. The newly introduced lookup dictionary incorporates rich contextual information in training set, which is vital to correctly predict long-tail tokens. Our proposed method is proved to outperform the baseline Transformer LM by a great margin on both word/character error rate and tail tokens error rate.
  arXiv  Detail & Related papers  (2022-12-30T22:26:57Z)
• Language Model Pre-Training with Sparse Latent Typing [66.75786739499604]
  We propose a new pre-training objective, Sparse Latent Typing, which enables the model to sparsely extract sentence-level keywords with diverse latent types. Experimental results show that our model is able to learn interpretable latent type categories in a self-supervised manner without using any external knowledge.
  arXiv  Detail & Related papers  (2022-10-23T00:37:08Z)
• An Interpretability Evaluation Benchmark for Pre-trained Language Models [37.16893581395874]
  We propose a novel evaluation benchmark providing with both English and Chinese annotated data. It tests LMs abilities in multiple dimensions, i.e., grammar, semantics, knowledge, reasoning and computation. It contains perturbed instances for each original instance, so as to use the rationale consistency under perturbations as the metric for faithfulness.
  arXiv  Detail & Related papers  (2022-07-28T08:28:09Z)

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.