Instruction-tuned Language Models are Better Knowledge Learners
- URL: http://arxiv.org/abs/2402.12847v2
- Date: Sun, 26 May 2024 03:19:48 GMT
- Title: Instruction-tuned Language Models are Better Knowledge Learners
- Authors: Zhengbao Jiang, Zhiqing Sun, Weijia Shi, Pedro Rodriguez, Chunting Zhou, Graham Neubig, Xi Victoria Lin, Wen-tau Yih, Srinivasan Iyer,
- Abstract summary: We propose pre-instruction-tuning (PIT), a method that instruction-tunes on questions prior to training on documents.
Extensive experiments and ablation studies demonstrate that pre-instruction-tuning significantly enhances the ability of LLMs to absorb knowledge from new documents.
- Score: 106.38526595116961
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: In order for large language model (LLM)-based assistants to effectively adapt to evolving information needs, it must be possible to update their factual knowledge through continued training on new data. The standard recipe for doing so involves continued pre-training on new documents followed by instruction-tuning on question-answer (QA) pairs. However, we find that LLMs trained with this recipe struggle to answer questions, even though the perplexity of documents is minimized. We found that QA pairs are generally straightforward, while documents are more complex, weaving many factual statements together in an intricate manner. Therefore, we hypothesize that it is beneficial to expose LLMs to QA pairs before continued pre-training on documents so that the process of encoding knowledge from complex documents takes into account how this knowledge is accessed through questions. Based on this, we propose pre-instruction-tuning (PIT), a method that instruction-tunes on questions prior to training on documents. This contrasts with standard instruction-tuning, which learns how to extract knowledge after training on documents. Extensive experiments and ablation studies demonstrate that pre-instruction-tuning significantly enhances the ability of LLMs to absorb knowledge from new documents, outperforming standard instruction-tuning by 17.8%.
Related papers
- Self-Tuning: Instructing LLMs to Effectively Acquire New Knowledge through Self-Teaching [67.11497198002165]
Large language models (LLMs) often struggle to provide up-to-date information due to their one-time training.
Motivated by the remarkable success of the Feynman Technique in efficient human learning, we introduce Self-Tuning.
arXiv Detail & Related papers (2024-06-10T14:42:20Z) - R4: Reinforced Retriever-Reorder-Responder for Retrieval-Augmented Large Language Models [32.598670876662375]
Retrieval-augmented large language models (LLMs) leverage relevant content retrieved by information retrieval systems to generate correct responses.
Existing retriever-responder methods typically append relevant documents to the prompt of LLMs to perform text generation tasks.
We propose a new pipeline named "Reinforced Retriever-Reorder-Responder" to learn document orderings for retrieval-augmented LLMs.
arXiv Detail & Related papers (2024-05-04T12:59:10Z) - Source-Aware Training Enables Knowledge Attribution in Language Models [81.13048060332775]
Large language models (LLMs) learn a vast amount of knowledge during pretraining, but they are often oblivious to the source(s) of such knowledge.
We investigate the problem of intrinsic source citation, where LLMs are required to cite the pretraining source supporting a generated response.
Our training recipe can enable faithful attribution to the pretraining data without a substantial impact on the model's quality compared to standard pretraining.
arXiv Detail & Related papers (2024-04-01T09:39:38Z) - KIWI: A Dataset of Knowledge-Intensive Writing Instructions for
Answering Research Questions [63.307317584926146]
Large language models (LLMs) adapted to follow user instructions are now widely deployed as conversational agents.
In this work, we examine one increasingly common instruction-following task: providing writing assistance to compose a long-form answer.
We construct KIWI, a dataset of knowledge-intensive writing instructions in the scientific domain.
arXiv Detail & Related papers (2024-03-06T17:16:44Z) - BIDER: Bridging Knowledge Inconsistency for Efficient Retrieval-Augmented LLMs via Key Supporting Evidence [23.55601157586831]
This paper introduces BIDER, an approach that refines retrieval documents into Key Supporting Evidence.
We train BIDER by learning from crafting KSE, while maximizing its output to align with LLM's information acquisition preferences.
Evaluations show BIDER boosts LLMs' answer quality by 7% while reducing input content length in retrieval documents by 80%, outperforming existing methods.
arXiv Detail & Related papers (2024-02-19T14:28:31Z) - Where is the answer? Investigating Positional Bias in Language Model Knowledge Extraction [36.40833517478628]
Large language models require updates to remain up-to-date or adapt to new domains.
One key is memorizing the latest information in a way that the memorized information is extractable with a query prompt.
Despite minimizing document perplexity during fine-tuning, LLMs struggle to extract information through a prompt sentence.
arXiv Detail & Related papers (2024-02-16T06:29:16Z) - Improving Information Extraction on Business Documents with Specific
Pre-Training Tasks [1.9331361036118608]
Transformer-based Language Models are widely used in Natural Language Processing related tasks.
We introduce two new pre-training tasks that force the model to learn better-contextualized representations of the scanned documents.
We also introduce a new post-processing algorithm to decode BIESO tags in Information Extraction that performs better with complex entities.
arXiv Detail & Related papers (2023-09-11T13:05:23Z) - Eva-KELLM: A New Benchmark for Evaluating Knowledge Editing of LLMs [54.22416829200613]
Eva-KELLM is a new benchmark for evaluating knowledge editing of large language models.
Experimental results indicate that the current methods for knowledge editing using raw documents are not effective in yielding satisfactory results.
arXiv Detail & Related papers (2023-08-19T09:17:19Z) - Knowledgeable Salient Span Mask for Enhancing Language Models as
Knowledge Base [51.55027623439027]
We develop two solutions to help the model learn more knowledge from unstructured text in a fully self-supervised manner.
To our best knowledge, we are the first to explore fully self-supervised learning of knowledge in continual pre-training.
arXiv Detail & Related papers (2022-04-17T12:33:34Z) - REALM: Retrieval-Augmented Language Model Pre-Training [37.3178586179607]
We augment language model pre-training with a latent knowledge retriever, which allows the model to retrieve and attend over documents from a large corpus such as Wikipedia.
For the first time, we show how to pre-train such a knowledge retriever in an unsupervised manner.
We demonstrate the effectiveness of Retrieval-Augmented Language Model pre-training (REALM) by fine-tuning on the challenging task of Open-domain Question Answering (Open-QA)
arXiv Detail & Related papers (2020-02-10T18:40:59Z)
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.