How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources

• URL: http://arxiv.org/abs/2306.04751v2
• Date: Mon, 30 Oct 2023 20:36:20 GMT
• Title: How Far Can Camels Go? Exploring the State of Instruction Tuning on Open Resources
• Authors: Yizhong Wang, Hamish Ivison, Pradeep Dasigi, Jack Hessel, Tushar Khot, Khyathi Raghavi Chandu, David Wadden, Kelsey MacMillan, Noah A. Smith, Iz Beltagy, Hannaneh Hajishirzi
• Abstract summary: This work explores recent advances in instruction-tuning language models on a range of open instruction-following datasets. We provide a large set of instruction-tuned models from 6.7B to 65B parameters in size, trained on 12 instruction datasets. We evaluate them on their factual knowledge, reasoning, multilinguality, coding, and open-ended instruction following abilities.
• Score: 117.6496550359768
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In this work we explore recent advances in instruction-tuning language models on a range of open instruction-following datasets. Despite recent claims that open models can be on par with state-of-the-art proprietary models, these claims are often accompanied by limited evaluation, making it difficult to compare models across the board and determine the utility of various resources. We provide a large set of instruction-tuned models from 6.7B to 65B parameters in size, trained on 12 instruction datasets ranging from manually curated (e.g., OpenAssistant) to synthetic and distilled (e.g., Alpaca) and systematically evaluate them on their factual knowledge, reasoning, multilinguality, coding, and open-ended instruction following abilities through a collection of automatic, model-based, and human-based metrics. We further introduce T\"ulu, our best performing instruction-tuned model suite finetuned on a combination of high-quality open resources. Our experiments show that different instruction-tuning datasets can uncover or enhance specific skills, while no single dataset (or combination) provides the best performance across all evaluations. Interestingly, we find that model and human preference-based evaluations fail to reflect differences in model capabilities exposed by benchmark-based evaluations, suggesting the need for the type of systemic evaluation performed in this work. Our evaluations show that the best model in any given evaluation reaches on average 87% of ChatGPT performance, and 73% of GPT-4 performance, suggesting that further investment in building better base models and instruction-tuning data is required to close the gap. We release our instruction-tuned models, including a fully finetuned 65B T\"ulu, along with our code, data, and evaluation framework at https://github.com/allenai/open-instruct to facilitate future research.

Related papers

• LIMIT: Less Is More for Instruction Tuning Across Evaluation Paradigms [2.249916681499244]
  We finetune open-source MPT-7B and MPT-30B models on instruction finetuning datasets of various sizes ranging from 1k to 60k samples. We find that subsets of 1k-6k instruction finetuning samples are sufficient to achieve good performance on both (1) traditional NLP benchmarks and (2) model-based evaluation.
  arXiv  Detail & Related papers  (2023-11-22T03:37:01Z)
• Fantastic Gains and Where to Find Them: On the Existence and Prospect of General Knowledge Transfer between Any Pretrained Model [74.62272538148245]
  We show that for arbitrary pairings of pretrained models, one model extracts significant data context unavailable in the other. We investigate if it is possible to transfer such "complementary" knowledge from one model to another without performance degradation.
  arXiv  Detail & Related papers  (2023-10-26T17:59:46Z)
• Harnessing the Power of David against Goliath: Exploring Instruction Data Generation without Using Closed-Source Models [32.41573520305861]
  We explore alternative approaches to generate high-quality instruction data that do not rely on closed-source models. Evaluation results from two benchmarks and the GPT-4 model demonstrate the effectiveness of our generated instruction data.
  arXiv  Detail & Related papers  (2023-08-24T11:07:47Z)
• INSTRUCTEVAL: Towards Holistic Evaluation of Instruction-Tuned Large Language Models [39.46610170563634]
  INSTRUCTEVAL is a more comprehensive evaluation suite designed specifically for instruction-tuned large language models. We take a holistic approach to analyze various factors affecting model performance, including the pretraining foundation, instruction-tuning data, and training methods. Our findings reveal that the quality of instruction data is the most crucial factor in scaling model performance.
  arXiv  Detail & Related papers  (2023-06-07T20:12:29Z)
• Universal Domain Adaptation from Foundation Models: A Baseline Study [58.51162198585434]
  We make empirical studies of state-of-the-art UniDA methods using foundation models. We introduce textitCLIP distillation, a parameter-free method specifically designed to distill target knowledge from CLIP models. Although simple, our method outperforms previous approaches in most benchmark tasks.
  arXiv  Detail & Related papers  (2023-05-18T16:28:29Z)
• Dataless Knowledge Fusion by Merging Weights of Language Models [51.8162883997512]
  Fine-tuning pre-trained language models has become the prevalent paradigm for building downstream NLP models. This creates a barrier to fusing knowledge across individual models to yield a better single model. We propose a dataless knowledge fusion method that merges models in their parameter space.
  arXiv  Detail & Related papers  (2022-12-19T20:46:43Z)
• Scaling Instruction-Finetuned Language Models [126.4789306516927]
  Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance. We find that instruction finetuning dramatically improves performance on a variety of model classes.
  arXiv  Detail & Related papers  (2022-10-20T16:58:32Z)
• ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Models [102.63817106363597]
  We build ELEVATER, the first benchmark to compare and evaluate pre-trained language-augmented visual models. It consists of 20 image classification datasets and 35 object detection datasets, each of which is augmented with external knowledge. We will release our toolkit and evaluation platforms for the research community.
  arXiv  Detail & Related papers  (2022-04-19T10:23:42Z)
• Bringing Light Into the Dark: A Large-scale Evaluation of Knowledge Graph Embedding Models Under a Unified Framework [31.35912529064612]
  We re-implemented and evaluated 21 interaction models in the PyKEEN software package. We performed a large-scale benchmarking on four datasets with several thousands of experiments and 24,804 GPU hours of time.
  arXiv  Detail & Related papers  (2020-06-23T22:30:52Z)
• Learning to Compare for Better Training and Evaluation of Open Domain Natural Language Generation Models [23.62054164511058]
  We propose to evaluate natural language generation models by learning to compare a pair of generated sentences by fine-tuning BERT. While able to be trained in a fully self-supervised fashion, our model can be further fine-tuned with a little amount of human preference annotation.
  arXiv  Detail & Related papers  (2020-02-12T15:52:21Z)

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.