ProsocialDialog: A Prosocial Backbone for Conversational Agents

Most existing dialogue systems fail to respond properly to potentially unsafe user utterances by either ignoring or passively agreeing with them. To address this issue, we introduce ProsocialDialog, the first large-scale multi-turn dialogue dataset to teach conversational agents to respond to problematic content following social norms. Covering diverse unethical, problematic, biased, and toxic situations, ProsocialDialog contains responses that encourage prosocial behavior, grounded in commonsense social rules (i.e., rules-of-thumb, RoTs). Created via a human-AI collaborative framework, ProsocialDialog consists of 58K dialogues, with 331K utterances, 160K RoTs, and 497K dialogue safety labels accompanied by free-form rationales. With this dataset, we introduce a dialogue safety detection module, Canary, capable of generating RoTs given conversational context, and a socially-informed dialogue agent, Prost. Empirical results show that Prost generates more socially acceptable dialogues compared to other state-of-the-art language and dialogue models in both in-domain and out-of-domain settings. Additionally, Canary effectively guides conversational agents and off-the-shelf language models to generate significantly more prosocial responses. Our work highlights the promise and importance of creating and steering conversational AI to be socially responsible.

Aligning to Social Norms and Values in Interactive Narratives

We focus on creating agents that act in alignment with socially beneficial norms and values in interactive narratives or text-based games -- environments wherein an agent perceives and interacts with a world through natural language. Such interactive agents are often trained via reinforcement learning to optimize task performance, even when such rewards may lead to agent behaviors that violate societal norms -- causing harm either to the agent itself or other entities in the environment. Social value alignment refers to creating agents whose behaviors conform to expected moral and social norms for a given context and group of people -- in our case, it means agents that behave in a manner that is less harmful and more beneficial for themselves and others. We build on the Jiminy Cricket benchmark (Hendrycks et al. 2021), a set of 25 annotated interactive narratives containing thousands of morally salient scenarios covering everything from theft and bodily harm to altruism. We introduce the GALAD (Game-value ALignment through Action Distillation) agent that uses the social commonsense knowledge present in specially trained language models to contextually restrict its action space to only those actions that are aligned with socially beneficial values. An experimental study shows that the GALAD agent makes decisions efficiently enough to improve state-of-the-art task performance by 4% while reducing the frequency of socially harmful behaviors by 25% compared to strong contemporary value alignment approaches.

ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection

Toxic language detection systems often falsely flag text that contains minority group mentions as toxic, as those groups are often the targets of online hate. Such over-reliance on spurious correlations also causes systems to struggle with detecting implicitly toxic language. To help mitigate these issues, we create ToxiGen, a new large-scale and machine-generated dataset of 274k toxic and benign statements about 13 minority groups. We develop a demonstration-based prompting framework and an adversarial classifier-in-the-loop decoding method to generate subtly toxic and benign text with a massive pretrained language model. Controlling machine generation in this way allows ToxiGen to cover implicitly toxic text at a larger scale, and about more demographic groups, than previous resources of human-written text. We conduct a human evaluation on a challenging subset of ToxiGen and find that annotators struggle to distinguish machine-generated text from human-written language. We also find that 94.5% of toxic examples are labeled as hate speech by human annotators. Using three publicly-available datasets, we show that finetuning a toxicity classifier on our data improves its performance on human-written data substantially. We also demonstrate that ToxiGen can be used to fight machine-generated toxicity as finetuning improves the classifier significantly on our evaluation subset.

Computational Lens on Cognition: Study Of Autobiographical Versus Imagined Stories With Large-Scale Language Models

Lifelong experiences and learned knowledge lead to shared expectations about how common situations tend to unfold. Such knowledge enables people to interpret story narratives and identify salient events effortlessly. We study differences in the narrative flow of events in autobiographical versus imagined stories using GPT-3, one of the largest neural language models created to date. The diary-like stories were written by crowdworkers about either a recently experienced event or an imagined event on the same topic. To analyze the narrative flow of events of these stories, we measured sentence *sequentiality*, which compares the probability of a sentence with and without its preceding story context. We found that imagined stories have higher sequentiality than autobiographical stories, and that the sequentiality of autobiographical stories is higher when they are retold than when freshly recalled. Through an annotation of events in story sentences, we found that the story types contain similar proportions of major salient events, but that the autobiographical stories are denser in factual minor events. Furthermore, in comparison to imagined stories, autobiographical stories contain more concrete words and words related to the first person, cognitive processes, time, space, numbers, social words, and core drives and needs. Our findings highlight the opportunity to investigate memory and cognition with large-scale statistical language models.

Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection

The perceived toxicity of language can vary based on someone's identity and beliefs, but this variation is often ignored when collecting toxic language datasets, resulting in dataset and model biases. We seek to understand the who, why, and what behind biases in toxicity annotations. In two online studies with demographically and politically diverse participants, we investigate the effect of annotator identities (who) and beliefs (why), drawing from social psychology research about hate speech, free speech, racist beliefs, political leaning, and more. We disentangle what is annotated as toxic by considering posts with three characteristics: anti-Black language, African American English (AAE) dialect, and vulgarity. Our results show strong associations between annotator identity and beliefs and their ratings of toxicity. Notably, more conservative annotators and those who scored highly on our scale for racist beliefs were less likely to rate anti-Black language as toxic, but more likely to rate AAE as toxic. We additionally present a case study illustrating how a popular toxicity detection system's ratings inherently reflect only specific beliefs and perspectives. Our findings call for contextualizing toxicity labels in social variables, which raises immense implications for toxic language annotation and detection.

Delphi: Towards Machine Ethics and Norms

What would it take to teach a machine to behave ethically? While broad ethical rules may seem straightforward to state ("thou shalt not kill"), applying such rules to real-world situations is far more complex. For example, while "helping a friend" is generally a good thing to do, "helping a friend spread fake news" is not. We identify four underlying challenges towards machine ethics and norms: (1) an understanding of moral precepts and social norms; (2) the ability to perceive real-world situations visually or by reading natural language descriptions; (3) commonsense reasoning to anticipate the outcome of alternative actions in different contexts; (4) most importantly, the ability to make ethical judgments given the interplay between competing values and their grounding in different contexts (e.g., the right to freedom of expression vs. preventing the spread of fake news). Our paper begins to address these questions within the deep learning paradigm. Our prototype model, Delphi, demonstrates strong promise of language-based commonsense moral reasoning, with up to 92.1% accuracy vetted by humans. This is in stark contrast to the zero-shot performance of GPT-3 of 52.3%, which suggests that massive scale alone does not endow pre-trained neural language models with human values. Thus, we present Commonsense Norm Bank, a moral textbook customized for machines, which compiles 1.7M examples of people's ethical judgments on a broad spectrum of everyday situations. In addition to the new resources and baseline performances for future research, our study provides new insights that lead to several important open research questions: differentiating between universal human values and personal values, modeling different moral frameworks, and explainable, consistent approaches to machine ethics.

Just Say No: Analyzing the Stance of Neural Dialogue Generation in Offensive Contexts

Dialogue models trained on human conversations inadvertently learn to generate offensive responses. Moreover, models can insult anyone by agreeing with an offensive context. To understand the dynamics of contextually offensive language, we study the stance of dialogue model responses in offensive Reddit conversations. Specifically, we crowd-annotate ToxiChat, a new dataset of 2,000 Reddit threads and model responses labeled with offensive language and stance. Our analysis reveals that 42% of user responses agree with toxic comments; 3x their agreement with safe comments (13%). Pre-trained transformer-based classifiers fine-tuned on our dataset achieve 0.71 F1 for offensive labels and 0.53 Macro-F1 for stance labels. Finally, we analyze some existing controllable text generation (CTG) methods to mitigate the contextual offensive behavior of dialogue models. Compared to the baseline, our best CTG model obtains a 19% reduction in agreement with offensive context and 29% fewer offensive responses. This highlights the need for future work to characterize and analyze more forms of inappropriate behavior in dialogue models to help make them safer. Our code and corpus are available at https://github.com/abaheti95/ToxiChat .

DExperts: Decoding-Time Controlled Text Generation with Experts and Anti-Experts

Despite recent advances in natural language generation, it remains challenging to control attributes of generated text. We propose DExperts: Decoding-time Experts, a decoding-time method for controlled text generation that combines a pretrained language model with "expert" LMs and/or "anti-expert" LMs in a product of experts. Intuitively, under the ensemble, tokens only get high probability if they are considered likely by the experts, and unlikely by the anti-experts. We apply DExperts to language detoxification and sentiment-controlled generation, where we outperform existing controllable generation methods on both automatic and human evaluations. Moreover, because DExperts operates only on the output of the pretrained LM, it is effective with (anti-)experts of smaller size, including when operating on GPT-3. Our work highlights the promise of tuning small LMs on text with (un)desirable attributes for efficient decoding-time steering.

On-the-Fly Controlled Text Generation with Experts and Anti-Experts

Despite recent advances in natural language generation, it remains challenging to control attributes of generated text. We propose DExperts: Decoding-time Experts, a decoding-time method for controlled text generation which combines a pretrained language model with experts and/or anti-experts in an ensemble of language models. Intuitively, under our ensemble, output tokens only get high probability if they are considered likely by the experts, and unlikely by the anti-experts. We apply DExperts to language detoxification and sentiment-controlled generation, where we outperform existing controllable generation methods on both automatic and human evaluations. Our work highlights the promise of using LMs trained on text with (un)desired attributes for efficient decoding-time controlled language generation.

Misinfo Belief Frames: A Case Study on Covid & Climate News

Prior beliefs of readers impact the way in which they project meaning onto news headlines. These beliefs can influence their perception of news reliability, as well as their reaction to news, and their likelihood of spreading the misinformation through social networks. However, most prior work focuses on fact-checking veracity of news or stylometry rather than measuring impact of misinformation. We propose Misinfo Belief Frames, a formalism for understanding how readers perceive the reliability of news and the impact of misinformation. We also introduce the Misinfo Belief Frames (MBF) corpus, a dataset of 66k inferences over 23.5k headlines. Misinformation frames use commonsense reasoning to uncover implications of real and fake news headlines focused on global crises: the Covid-19 pandemic and climate change. Our results using large-scale language modeling to predict misinformation frames show that machine-generated inferences can influence readers' trust in news headlines (readers' trust in news headlines was affected in 29.3% of cases). This demonstrates the potential effectiveness of using generated frames to counter misinformation.