An official website of the United States government
Large Language Models (LLMs) have achieved remarkable success in natural language tasks, yet understanding their reasoning processes re- mains a significant challenge. We address this by introducing XplainLLM, a dataset accom- panying an explanation framework designed to enhance LLM transparency and reliability. Our dataset comprises 24,204 instances where each instance interprets the LLM’s reasoning behavior using knowledge graphs (KGs) and graph attention networks (GAT), and includes explanations of LLMs such as the decoder- only Llama-3 and the encoder-only RoBERTa. XplainLLM also features a framework for gener- ating grounded explanations and the debugger- scores for multidimensional quality analysis. Our explanations include why-choose and why- not-choose components, reason-elements, and debugger-scores that collectively illuminate the LLM’s reasoning behavior. Our evaluations demonstrate XplainLLM’s potential to reduce hallucinations and improve grounded explana- tion generation in LLMs. XplainLLM is a re- source for researchers and practitioners to build trust and verify the reliability of LLM outputs. Our code and dataset are publicly available.
more »
« less
Explanations that reveal all through the definition of encoding
Feature attributions attempt to highlight what inputs drive predictive power. Good attributions or explanations are thus those that produce inputs that retain this predictive power; accordingly, evaluations of explanations score their quality of prediction. However, evaluations produce scores better than what appears possible from the values in the explanation for a class of explanations, called encoding explanations. Probing for encoding remains a challenge because there is no general characterization of what gives the extra predictive power. We develop a definition of encoding that identifies this extra predictive power via conditional dependence and show that the definition fits existing examples of encoding. This definition implies, in contrast to encoding explanations, that non-encoding explanations contain all the informative inputs used to produce the explanation, giving them a "what you see is what you get" property, which makes them transparent and simple to use. Next, we prove that existing scores (ROAR, FRESH, EVAL-X) do not rank non-encoding explanations above encoding ones, and develop STRIPE-X which ranks them correctly. After empirically demonstrating the theoretical insights, we use STRIPE-X to show that despite prompting an LLM to produce non-encoding explanations for a sentiment analysis task, the LLM-generated explanations encode.
more »
« less
- Award ID(s):
- 1922658
- PAR ID:
- 10649815
- Publisher / Repository:
- 38th Conference on Neural Information Processing Systems (NeurIPS 2024).
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Probabilistic graphical models, such as Markov random fields (MRF), exploit dependencies among random variables to model a rich family of joint probability distributions. Inference algorithms, such as belief propagation (BP), can effectively compute the marginal posteriors for decision making. Nonetheless, inferences involve sophisticated probability calculations and are difficult for humans to interpret. Among all existing explanation methods for MRFs, no method is designed for fair attributions of an inference outcome to elements on the MRF where the inference takes place. Shapley values provide rigorous attributions but so far have not been studied on MRFs. We thus define Shapley values for MRFs to capture both probabilistic and topological contributions of the variables on MRFs. We theoretically characterize the new definition regarding independence, equal contribution, additivity, and submodularity. As brute-force computation of the Shapley values is challenging, we propose GraphShapley, an approximation algorithm that exploits the decomposability of Shapley values, the structure of MRFs, and the iterative nature of BP inference to speed up the computation. In practice, we propose meta-explanations to explain the Shapley values and make them more accessible and trustworthy to human users. On four synthetic and nine real-world MRFs, we demonstrate that GraphShapley generates sensible and practical explanations.more » « less
-
With the increasing popularity of Graph Neural Networks (GNNs) for predictive tasks on graph structured data, research on their explainability is becoming more critical and achieving significant progress. Although many methods are proposed to explain the predictions of GNNs, their focus is mainly on “how to generate explanations.” However, other important research questions like “whether the GNN explanations are inaccurate,” “what if the explanations are inaccurate,” and “how to adjust the model to generate more accurate explanations” have gained little attention. Our previous GNN Explanation Supervision (GNES) framework demonstrated effectiveness on improving the reasonability of the local explanation while still keep or even improve the backbone GNNs model performance. In many applications instead of per sample explanations, we need to find global explanations which are reasonable and faithful to the domain data. Simply learning to explain GNNs locally is not an optimal solution to a global understanding of the model. To improve the explainability power of the GNES framework, we propose the Global GNN Explanation Supervision (GGNES) technique which uses a basic trained GNN and a global extension of the loss function used in the GNES framework. This GNN creates local explanations which are fed to a Global Logic-based GNN Explainer, an existing technique that can learn the global Explanation in terms of a logic formula. These two frameworks are then trained iteratively to generate reasonable global explanations. Extensive experiments demonstrate the effectiveness of the proposed model on improving the global explanations while keeping the performance similar or even increase the model prediction power.more » « less
-
Large language models (LLMs) have revolution- ized machine learning due to their ability to cap- ture complex interactions between input features. Popular post-hoc explanation methods like SHAP provide marginal feature attributions, while their extensions to interaction importances only scale to small input lengths (≈20). We propose Spectral Ex- plainer (SPEX), a model-agnostic interaction attri- bution algorithm that efficiently scales to large input lengths (≈1000). SPEX exploits underlying nat- ural sparsity among interactions—common in real- world data—and applies a sparse Fourier transform using a channel decoding algorithm to efficiently identify important interactions. We perform exper- iments across three difficult long-context datasets that require LLMs to utilize interactions between inputs to complete the task. For large inputs, SPEX outperforms marginal attribution methods by up to 20% in terms of faithfully reconstructing LLM out- puts. Further, SPEX successfully identifies key fea- tures and interactions that strongly influence model output. For one of our datasets, HotpotQA, SPEX provides interactions that align with human annota- tions. Finally, we use our model-agnostic approach to generate explanations to demonstrate abstract rea- soning in closed-source LLMs (GPT-4o mini) and compositional reasoning in vision-language models.more » « less
-
This paper systematically investigates the generation of code explanations by Large Language Models (LLMs) for code examples commonly encountered in introductory programming courses. Our findings reveal significant variations in the nature of code explanations produced by LLMs, influenced by factors such as the wording of the prompt, the specific code examples under consideration, the programming language involved, the temperature parameter, and the version of the LLM. However, a consistent pattern emerges for Java and Python, where explanations exhibit a Flesch-Kincaid readability level of approximately 7-8 grade and a consistent lexical density, indicating the proportion of meaningful words relative to the total explanation size. Additionally, the generated explanations consistently achieve high scores for correctness, but lower scores on three other metrics: completeness, conciseness, and specificity.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
