Challenge: a little or a large bias in CtB-LLMs may cause huge harm . LLaMA and OPT families have an important bias in gender, race, religion, and profession.
Approach: They propose to debiase three families of Very Large-Language Models with LORA to reduce bias by 4.12 points in the normalized stereotype score.
Outcome: The proposed model reduces bias up to 4.12 points in the normalized stereotype score.

Similar Papers

A Generative Approach for Mitigating Structural Biases in Natural Language Inference (2022.starsem-1)

Copied to clipboard

Challenge: Natural language inference datasets contain artifacts and biases that allow models to perform poorly by using a biased subset of the input without considering the remainder features.
Approach: They reformulate a natural language inference task as a generative task . they find that this approach is highly robust to large amounts of bias .
Outcome: The proposed model is highly robust to large amounts of bias.
LEXPLAIN: Improving Model Explanations via Lexicon Supervision (2023.starsem-1)

Copied to clipboard

Challenge: Existing methods that extract features from input text to explain a classifier's prediction are limiting to models that are faithful to their predictions.
Approach: They propose a framework for guiding model explanations by supervising them explicitly using task-related lexicons to direct supervise model explanation.
Outcome: The proposed method improves model explanations without sacrificing performance on sentiment analysis and toxicity detection tasks while demoting spurious correlations with African American English dialects.
Language models are not naysayers: an analysis of language models on negation benchmarks (2023.starsem-1)

Copied to clipboard

Challenge: Negation has been shown to be a major bottleneck for masked language models, such as BERT, but whether this finding still holds for larger-sized auto-regressive language models has not been studied comprehensively.
Approach: They evaluate the ability of current-generation auto-regressive language models to handle negation using a wide range of benchmarks and models.
Outcome: The proposed models are compared against a wide range of negation benchmarks and show that they are insensitive to negation, inability to capture the lexical semantics of negations, and failure to reason under negation.
Natural Language Inference with Mixed Effects (2020.starsem-1)

Copied to clipboard

Challenge: aggregating raw annotations to a single label is problematic due to disagreement among annotators.
Approach: They propose a generic method that allows one to skip the aggregation step and train on the raw annotations directly without subjecting the model to unwanted noise.
Outcome: The proposed method improves performance over models that do not incorporate such effects.
Exploring Factual Entailment with NLI: A News Media Study (2024.starsem-1)

Copied to clipboard

Challenge: Recent studies have focused on the relationship between factuality and Natural Language Inference (NLI).
Approach: They propose a novel annotation scheme that models factual rather than textual entailment and use it to annotate a dataset of naturally occurring sentences from news articles.
Outcome: The proposed annotation scheme can be used to model factual relationships on a dataset of naturally occurring sentences from news articles.
Syntax and Semantics Meet in the “Middle”: Probing the Syntax-Semantics Interface of LMs Through Agentivity (2023.starsem-1)

Copied to clipboard

Challenge: a recent study examined how large language models handle interactions in meaning across words and larger syntactic forms.
Approach: They propose to use a dataset to examine the linguistic properties of optionally transitive English verbs to examine their agentivity.
Outcome: The proposed model outperforms all other models in the evaluation dataset . the results are better correlated with human judgements than syntactic and semantic corpus statistics .
Limits for learning with language models (2023.starsem-1)

Copied to clipboard

Challenge: Recent studies show that large language models fail to capture important aspects of linguistic meaning . authors argue that LLMs cannot learn fundamental semantic properties defined in formal semantics .
Approach: They propose a theoretical explanation for some of the observed failings of large language models . they show that LLMs cannot learn certain fundamental semantic properties .
Outcome: The proposed model fails to learn semantic entailment and consistency as defined in formal semantics, the authors argue . their model fails on tasks that require engorgements and deep linguistic understanding, they argue - but not on universal quantification.
How Does Stereotype Content Differ across Data Sources? (2024.starsem-1)

Copied to clipboard

Challenge: Existing studies of stereotypes using rating scales capture beliefs and opinions about different social groups.
Approach: They compare stereotype-relevant measures of social group social status with traditional scales and a word-list generation task using free-text data.
Outcome: The results compare with traditional surveys and a spontaneous word-list generation task.
How Does Data Corruption Affect Natural Language Understanding Models? A Study on GLUE datasets (2022.starsem-1)

Copied to clipboard

Challenge: Existing studies on the performance of pre-trained language models on natural language understanding tasks have focused on the natural language inference and textual entailment tasks.
Approach: They propose to use corrupted data to fine-tune pre-trained language models to assess their language understanding capabilities.
Outcome: The proposed transformations can be applied to all but one NLU task and show that understanding the meaning of utterances is not required for high performance.
Analyzing Syntactic Generalization Capacity of Pre-trained Language Models on Japanese Honorific Conversion (2023.starsem-1)

Copied to clipboard

Challenge: Using Japanese honorifics requires knowledge of grammatical rules and contextual information, such as social relationships.
Approach: They propose a Japanese honorific conversion task that considers social relationships among people mentioned in a conversation.
Outcome: The proposed model performed better on the context-aware task than the prompt-based one.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations