Reframing Human-AI Collaboration for Generating Free-Text Explanations

Wiegreffe, Sarah; Hessel, Jack; Swayamdipta, Swabha; Riedl, Mark O.; Choi, Yejin

doi:10.18653/v1/2022.naacl-main.47

Cited by 34 publications

(39 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Concurrent to our work, Yordanov et al ( 2021) study self-rationalization transfer from a highresource task to a task with only a few humanauthored explanations. Wiegreffe et al (2022) analyze explanations obtained by prompting GPT-3 multiple times to get multiple explanation candidates, and then filter these candidates using a model trained to predict acceptability of explanations. Their prompt consists of a few examples with high-quality explanations written by the authors and a new instance together with its gold label.…”

Section: Related Workmentioning

confidence: 99%

Few-Shot Self-Rationalization with Natural Language Prompts

Marasović¹,

Beltagy²,

Downey³

et al. 2022

Findings of the Association for Computational Linguistics: NAACL 2022

View full text Add to dashboard Cite

Self-rationalization models that predict task labels and generate free-text elaborations for their predictions could enable more intuitive interaction with NLP systems. These models are, however, currently trained with a large amount of human-written free-text explanations for each task which hinders their broader usage. We propose to study a more realistic setting of self-rationalization using few training examples. We present FEB-a standardized collection of four existing English-language datasets and associated metrics. We identify the right prompting approach by extensively exploring natural language prompts on FEB. Then, by using this prompt and scaling the model size, we demonstrate that making progress on few-shot self-rationalization is possible. We show there is still ample room for improvement in this task: the average plausibility of generated explanations assessed by human annotators is at most 51% (with GPT-3), while plausibility of human explanations is 76%. We hope that FEB and our proposed approach will spur the community to take on the few-shot self-rationalization challenge.

show abstract

Section: Related Workmentioning

confidence: 99%

Few-Shot Self-Rationalization with Natural Language Prompts

Marasović¹,

Beltagy²,

Downey³

et al. 2022

Findings of the Association for Computational Linguistics: NAACL 2022

View full text Add to dashboard Cite

show abstract

“…has shown that task performance can be improved by sampling multiple language model outputs for ensembling, (2) prompt-order ensembling, where previous work (Lu et al, 2021;Zhao et al, 2021) has shown that task performance is sensitive to the order of the exemplars in the prompts, and (3) input-rationale ensembling, where human-written rationales can be replaced by model-generated rationales, leveraging the ability of language models to generate high-quality explanations (Wiegreffe et al, 2022). Figure 1 provides an overview of rationale-augmented ensembling approaches.…”

Section: Language Modelmentioning

confidence: 99%

“…Rationale-augmented ensembles Input/Prompt Output Self-consistency (Wang et al, 2022) fixed sampled Prompt-order ensemble (Lu et al, 2021;Zhao et al, 2021) shuffled greedy/sampled Input-rationale ensemble, adapted from (Wiegreffe et al, 2022) sampled greedy/sampled Table 2: Methods for generating rationale-augmented ensembles in language models.…”

Section: Rationale-augmented Ensembles In Language Modelsmentioning

confidence: 99%

“…• Input-rationale ensemble: Leveraging the ability of large language models to generate highquality explanations (Wiegreffe et al, 2022), model-generated rationales can replace humanwritten rationales in the input prompts (e.g., via the process described in Section 2.1), which can then be used to elicit multiple rationales in the decoder.…”

Section: Rationale-augmented Ensemblesmentioning

confidence: 99%

See 1 more Smart Citation

Rationale-Augmented Ensembles in Language Models

Wang¹,

Lee²,

Schuurmans³

et al. 2022

Preprint

View full text Add to dashboard Cite

Recent research has shown that rationales, or step-by-step chains of thought, can be used to improve performance in multi-step reasoning tasks. We reconsider rationale-augmented prompting for few-shot in-context learning, where (input → output) prompts are expanded to (input, rationale → output) prompts. For rationale-augmented prompting we demonstrate how existing approaches, which rely on manual prompt engineering, are subject to sub-optimal rationales that may harm performance. To mitigate this brittleness, we propose a unified framework of rationale-augmented ensembles, where we identify rationale sampling in the output space as the key component to robustly improve performance. This framework is general and can easily be extended to common natural language processing tasks, even those that do not traditionally leverage intermediate steps, such as question answering, word sense disambiguation, and sentiment analysis. We demonstrate that rationale-augmented ensembles achieve more accurate and interpretable results than existing prompting approaches-including standard prompting without rationales and rationale-based chain-of-thought prompting-while simultaneously improving interpretability of model predictions through the associated rationales.Preprint.

show abstract

“…GPT-3 Rationales for Gold Labels. Wiegreffe et al (2022) collected 250 high quality free-text rationales generated by few-shot prompting with GPT-3 (Brown et al, 2020) for CQA (given gold labels). Each example was assessed by 3 crowdworkers.…”

Section: Evaluating Rationales In Few-shot Promptingmentioning

confidence: 99%

REV: Information-Theoretic Evaluation of Free-Text Rationales

Chen¹,

Brahman²,

Ren³

et al. 2022

Preprint

View full text Add to dashboard Cite

Free-text rationales are a promising step towards explainable AI, yet their evaluation remains an open research problem. While existing metrics have mostly focused on measuring the direct association between the rationale and a given label, we argue that an ideal metric should also be able to focus on the new information uniquely provided in the rationale that is otherwise not provided in the input or the label. We investigate this research problem from an information-theoretic perspective using the conditional V-information (Hewitt et al., 2021). More concretely, we propose a metric called REV (Rationale Evaluation with conditional V-information), that can quantify the new information in a rationale supporting a given label beyond the information already available in the input or the label. Experiments on reasoning tasks across four benchmarks, including few-shot prompting with GPT-3, demonstrate the effectiveness of REV in evaluating different types of rationale-label pairs, compared to existing metrics. Through several quantitative comparisons, we demonstrate the capability of REV in providing more sensitive measurements of new information in free-text rationales with respect to a label. Furthermore, REV is consistent with human judgments on rationale evaluations. Overall, when used alongside traditional performance metrics, REV provides deeper insights into a model's reasoning and prediction processes.

show abstract

Reframing Human-AI Collaboration for Generating Free-Text Explanations

Cited by 34 publications

References 24 publications

Few-Shot Self-Rationalization with Natural Language Prompts

Few-Shot Self-Rationalization with Natural Language Prompts

Rationale-Augmented Ensembles in Language Models

REV: Information-Theoretic Evaluation of Free-Text Rationales

Contact Info

Product

Resources

About