Better Explain Transformers by Illuminating Important Information (2024.findings-eacl)
Copied to clipboard
| Challenge: | Existing explanations focus on the input and output of the Transformers, resulting in confusing results. |
| Approach: | They propose to highlight important information and eliminate irrelevant information by a refined information flow on top of the layer-wise relevance propagation method. |
| Outcome: | The proposed method outperforms baseline models on classification and question-answering datasets with over 3% to 33% improvement on explanation metrics. |
Similar Papers
VISIT: Visualizing and Interpreting the Semantic Information Flow of Transformers (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Recent work in interpretability suggests we can project weights and hidden states of transformer-based language models (LMs) to their vocabulary space, a transformation that makes them more human interpretable. |
| Approach: | They propose a tool to visualize a forward pass of Generative Pre-trained Transformers as an interactive flow graph with nodes representing neurons or hidden states and edges representing interactions between them. |
| Outcome: | The proposed visualization simplifies huge amounts of data into easy-to-read graphs that can reflect the models’ internal processing, uncovering the contribution of each component to the models' final prediction. |
Incorporating Residual and Normalization Layers into Analysis of Masked Language Models (2021.emnlp-main)
Copied to clipboard
| Challenge: | Transformer architecture is composed of multi-head attention, which has been extensively analyzed. |
| Approach: | They extended the scope of the analysis of Transformers from solely the attention patterns to the whole attention block, i.e., multi-head attention, residual connection, and layer normalization. |
| Outcome: | The proposed method incorporates the whole attention block, i.e., multi-head attention, residual connection, and layer normalization into the analysis. |
Explaining Text Similarity in Transformer Models (2024.naacl-long)
Copied to clipboard
| Challenge: | Modern foundation models provide flexible text representations that enable the detection of semantic structure in vast amounts of unlabeled data. |
| Approach: | They propose to leverage layer-wise relevance propagation to understand the inner prediction mechanisms of NLP models by analyzing grammatical interactions, multilingual semantics, and biomedical text retrieval. |
| Outcome: | The proposed methods demonstrate their utility in three corpus-level use cases, analyzing grammatical interactions, multilingual semantics, and biomedical text retrieval. |
Quantifying Attention Flow in Transformers (2020.acl-main)
Copied to clipboard
| Challenge: | In the Transformer model, “self-attention” combines information from attended embeddings into the representation of the focal embeddable in the next layer. |
| Approach: | They propose two methods to quantify flow of information through self-attention using attention weights as relative relevance of input tokens. |
| Outcome: | The proposed methods give complementary views on the flow of information and yield higher correlations with importance scores of input tokens. |
Refining Attention for Explainable and Noise-Robust Fact-Checking with Transformers (2025.emnlp-main)
Copied to clipboard
| Challenge: | Conventional transformer-based models falter due to noise sensitivity and lack explainability . ATTUN is a transformer architecture designed to enhance model transparency and resilience to noise. |
| Approach: | They propose a transformer architecture that enhances model transparency and resilience to noise . ATTUN is a module that directly modifies attention weights . they validated their approach using fact-checking datasets based on their results . |
| Outcome: | The proposed model improves predictions and identify relevant sections of input data. |
On the weak link between importance and prunability of attention heads (2020.emnlp-main)
Copied to clipboard
| Challenge: | a large fraction of attention heads can be randomly pruned with limited effect on accuracy, a new study finds . a second study finds no advantage in pruning attention heads identified to be important based on the location of a head . |
| Approach: | They examine the importance of pruning attention heads on a Transformer-based model . they find no advantage in pruning attention head positions on the BERT model based on location . |
| Outcome: | The results show that pruning strategies on Transformer and BERT models are not important based on location . the results suggest that interpretation of attention heads does not strongly inform pruning strategies. |
Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions (2024.emnlp-main)
Copied to clipboard
| Challenge: | Using a method to identify next-token neurons, we find that some attention heads recognize contexts relevant to predicting a token and activate a downstream token-predicting neuron accordingly. |
| Approach: | They propose a method to identify next-token neurons and determine the upstream attention heads responsible for their activity in LLMs. |
| Outcome: | The proposed method identifies next-token neurons, finds prompts that highly activate them, and determines the upstream attention heads responsible. |
Interpretability-based Tailored Knowledge Editing in Transformers (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for modifying in-context learning fail to analyze the instability of in-constitu learning outcomes. |
| Approach: | They propose a model-based knowledge editing method that considers the unique information flow of each sample and aims to correct errors without costly retraining. |
| Outcome: | The proposed method exploits the critical role of feed-forward MLPs in decoder-only models and reveals diverse attribute recall across transformer layers, guiding edits to specific features at different depths and mitigating over-editing issues. |
Discovering and Mitigating Indirect Bias in Attention-Based Model Explanations (2024.findings-naacl)
Copied to clipboard
| Challenge: | Discrimination is the unfair treatment or prejudice directed towards individuals, groups, or certain ideas or beliefs, intentionally or unintentionally. |
| Approach: | They propose an algorithm to detect and mitigate indirect bias in transformer models by leveraging attention explanations. |
| Outcome: | The proposed algorithm shows that it is more accurate than traditional fairness metrics and that it can be used to mitigate bias in transformer models. |
Transformer-specific Interpretability (2024.eacl-tutorials)
Copied to clipboard
| Challenge: | Transformers are dominant play-ers in various scientific fields, but their inner workings remain opaque. |
| Approach: | This tutorial presents a trending approach to interpreting Transformers . it uses specific features of the Transformer architecture to quantify context- mixing interactions . |
| Outcome: | This tutorial aims to show how a new trending approach can be applied to Transformer-based models. |