Tracr-Injection: Distilling Algorithms into Pre-trained Language Models

• URL: http://arxiv.org/abs/2505.10719v3
• Date: Mon, 02 Jun 2025 15:23:49 GMT
• Title: Tracr-Injection: Distilling Algorithms into Pre-trained Language Models
• Authors: Tomás Vergara-Browne, Álvaro Soto,
• Abstract summary: We propose tracr-injection, a method that allows us to distill algorithms written in RASP directly into a pre-trained language model.<n>We show how our method creates an interpretable subspace within the model's residual stream, which can be decoded into the variables present in the RASP algorithm.
• Score: 2.532202013576547
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Motivated by the surge of large language models, there has been a push to formally characterize the symbolic abilities intrinsic to the transformer architecture. A programming language, called RASP, has been proposed, which can be directly compiled into transformer weights to implement these algorithms. However, the tasks that can be implemented in RASP are often uncommon to learn from natural unsupervised data, showing a mismatch between theoretical capabilities of the transformer architecture, and the practical learnability of these capabilities from unsupervised data. We propose tracr-injection, a method that allows us to distill algorithms written in RASP directly into a pre-trained language model. We showcase our method by injecting 3 different algorithms into a language model. We show how our method creates an interpretable subspace within the model's residual stream, which can be decoded into the variables present in the code of the RASP algorithm. Additionally, we found that the proposed method can improve out-of-distribution performance compared to our baseline, indicating that indeed a more symbolic mechanism is taking place in the inner workings of the model. We release the code used to run our experiments.

Related papers

• Algorithmic Capabilities of Random Transformers [49.73113518329544]
  We investigate what functions can be learned by randomly transformers in which only the embedding layers are optimized. We find that these random transformers can perform a wide range of meaningful algorithmic tasks. Our results indicate that some algorithmic capabilities are present in transformers even before these models are trained.
  arXiv  Detail & Related papers  (2024-10-06T06:04:23Z)
• Transformers meet Neural Algorithmic Reasoners [16.5785372289558]
  We propose a novel approach that combines the Transformer's language understanding with the robustness of graph neural network (GNN)-based neural algorithmic reasoners (NARs) We evaluate our resulting TransNAR model on CLRS-Text, the text-based version of the CLRS-30 benchmark, and demonstrate significant gains over Transformer-only models for algorithmic reasoning.
  arXiv  Detail & Related papers  (2024-06-13T16:42:06Z)
• Transparency at the Source: Evaluating and Interpreting Language Models With Access to the True Distribution [4.01799362940916]
  We present a setup for training, evaluating and interpreting neural language models, that uses artificial, language-like data. The data is generated using a massive probabilistic grammar, that is itself derived from a large natural language corpus. With access to the underlying true source, our results show striking differences and outcomes in learning dynamics between different classes of words.
  arXiv  Detail & Related papers  (2023-10-23T12:03:01Z)
• Supervised Pretraining Can Learn In-Context Reinforcement Learning [96.62869749926415]
  In this paper, we study the in-context learning capabilities of transformers in decision-making problems. We introduce and study Decision-Pretrained Transformer (DPT), a supervised pretraining method where the transformer predicts an optimal action. We find that the pretrained transformer can be used to solve a range of RL problems in-context, exhibiting both exploration online and conservatism offline.
  arXiv  Detail & Related papers  (2023-06-26T17:58:50Z)
• Transformers as Statisticians: Provable In-Context Learning with In-Context Algorithm Selection [88.23337313766353]
  This work first provides a comprehensive statistical theory for transformers to perform ICL. We show that transformers can implement a broad class of standard machine learning algorithms in context. A emphsingle transformer can adaptively select different base ICL algorithms.
  arXiv  Detail & Related papers  (2023-06-07T17:59:31Z)
• Transformers as Algorithms: Generalization and Implicit Model Selection in In-context Learning [23.677503557659705]
  In-context learning (ICL) is a type of prompting where a transformer model operates on a sequence of examples and performs inference on-the-fly. We treat the transformer model as a learning algorithm that can be specialized via training to implement-at inference-time-another target algorithm. We show that transformers can act as an adaptive learning algorithm and perform model selection across different hypothesis classes.
  arXiv  Detail & Related papers  (2023-01-17T18:31:12Z)
• Thinking Like Transformers [64.96770952820691]
  We propose a computational model for the transformer-encoder in the form of a programming language. We show how RASP can be used to program solutions to tasks that could conceivably be learned by a Transformer. We provide RASP programs for histograms, sorting, and Dyck-languages.
  arXiv  Detail & Related papers  (2021-06-13T13:04:46Z)
• Unsupervised Paraphrasing with Pretrained Language Models [85.03373221588707]
  We propose a training pipeline that enables pre-trained language models to generate high-quality paraphrases in an unsupervised setting. Our recipe consists of task-adaptation, self-supervision, and a novel decoding algorithm named Dynamic Blocking. We show with automatic and human evaluations that our approach achieves state-of-the-art performance on both the Quora Question Pair and the ParaNMT datasets.
  arXiv  Detail & Related papers  (2020-10-24T11:55:28Z)
• Exploring Software Naturalness through Neural Language Models [56.1315223210742]
  The Software Naturalness hypothesis argues that programming languages can be understood through the same techniques used in natural language processing. We explore this hypothesis through the use of a pre-trained transformer-based language model to perform code analysis tasks.
  arXiv  Detail & Related papers  (2020-06-22T21:56:14Z)

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.