Extractive Fact Decomposition for Interpretable Natural Language Inference in one Forward Pass (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent work in Natural Language Inference (NLI) uses atomic fact decomposition to enhance interpretability and robustness. |
| Approach: | They propose an encoder-only architecture that performs extractive atomic fact decomposition and interpretable inference without generative models. |
| Outcome: | The proposed architecture achieves competitive accuracy and improves robustness out of distribution and in adversarial settings over models based on extractive rationale supervision. |
Similar Papers
Atomic Inference for NLI with Generated Facts as Atoms (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing models that can provide accurate explanations are not interpretable, i.e. they do not reflect the inner workings of the model. |
| Approach: | They propose to use LLM-generated facts as atoms to make interpretable models that can be used to make accurate predictions for each component part of an input. |
| Outcome: | The proposed method outperforms existing methods on natural language understanding tasks with a multi-stage fact generation process and a training regime that incorporates the facts. |
Entropy Guided Extrapolative Decoding to Improve Factuality in Large Language Models (2025.coling-main)
Copied to clipboard
| Challenge: | Large language models (LLMs) exhibit impressive natural language capabilities but suffer from hallucination – generating content that does not align with realworld facts. |
| Approach: | They propose to extrapolate critical token probabilities beyond the last layer to improve decoding by manipulating the predicted distributions at inference time. |
| Outcome: | The proposed methods surpass state-of-the-art on multiple datasets by large margins. |
Faithful and Robust LLM-Driven Theorem Proving for NLI Explanations (2025.acl-long)
Copied to clipboard
| Challenge: | Recent work has shown that the interaction of large language models (LLMs) with theorem provers (TPs) can help verify and improve the validity of NLI explanations. |
| Approach: | They propose to use logical expressions to guide LLMs in generating structured proof sketches and to use them to improve their accuracy. |
| Outcome: | The proposed strategies improve autoformalisation, syntactic errors and explanation refinement over the state-of-the-art model. |
NLI under the Microscope: What Atomic Hypothesis Decomposition Reveals (2025.naacl-long)
Copied to clipboard
| Challenge: | Decomposing text into atomic propositions allows for finergrained inspection of text. |
| Approach: | They propose to decompose atomic propositions into atomic sub-problems that models must weigh when solving the overall problem. |
| Outcome: | The proposed method measures the inferential consistency of models and the diversity of examples in benchmark datasets. |
Improving Unsupervised Commonsense Reasoning Using Knowledge-Enabled Natural Language Inference (2021.findings-emnlp)
Copied to clipboard
| Challenge: | Recent methods based on pre-trained language models have shown strong supervised performance on commonsense reasoning. |
| Approach: | They propose to use a common framework to solve commonsense reasoning tasks using a dataset from NLI. |
| Outcome: | The proposed method achieves state-of-the-art unsupervised performance on two commonsense reasoning tasks. |
Towards Intrinsic Interpretability of Large Language Models: A Survey of Design Principles and Architectures (2026.acl-long)
Copied to clipboard
| Challenge: | Existing studies on explainable AI focus on post-hoc explanation methods that interpret trained models through external approximations. |
| Approach: | They propose to categorize existing approaches into five design paradigms: functional transparency, concept alignment, representational decomposability, explicit modularization, and latent sparsity induction. |
| Outcome: | The proposed approaches are categorized into five design paradigms: functional transparency, concept alignment, representational decomposability, explicit modularization, and latent sparsity induction. |
FactReasoner: A Probabilistic Approach to Long-Form Factuality Assessment for Large Language Models (2025.findings-emnlp)
Copied to clipboard
Radu Marinescu, Debarun Bhattacharjya, Junkyu Lee, Tigran T. Tchrakian, Javier Carnerero-Cano, Yufang Hou, Elizabeth M. Daly, Alessandra Pascale
| Challenge: | Large language models often fail to ensure factual accuracy of outputs thus limiting reliability in real-world applications. |
| Approach: | They propose a neuro-symbolic based factuality assessment framework that employs probabilistic reasoning to evaluate the truthfulness of long-form generated responses. |
| Outcome: | The proposed framework outperforms state-of-the-art prompt-based methods in factual accuracy and recall. |
Less for More: Enhanced Feedback-aligned Mixed LLMs for Molecule Caption Generation and Fine-Grained NLI Evaluation (2025.acl-long)
Copied to clipboard
| Challenge: | Recent trends have led to the use of multimodal models to learn molecular and linguistic representations, either in separate but coordinated spaces or in a common space. |
| Approach: | They propose a novel atomic-level evaluation method leveraging off-the-shelf Natural Language Inference (NLI) models for use in the unseen chemical domain. |
| Outcome: | The proposed method surpasses state-of-the-art models in the unseen chemical domain while relying on a granularity-based evaluation method. |
Plausible Extractive Rationalization through Semi-Supervised Entailment Signal (2024.findings-acl)
Copied to clipboard
| Challenge: | Abstract: Large language models are gaining widespread adoption in natural language processing tasks. |
| Approach: | They propose a semi-supervised approach to optimize for plausibility of extracted rationales by using a pre-trained natural language inference model and a supervised NLI predictor. |
| Outcome: | The proposed model outperforms unsupervised models by > 100% on a ERASER dataset. |
Comprehensiveness Metrics for Automatic Evaluation of Factual Recall in Text Generation (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) produce incomplete or selectively omit key information . omissions of key information or misrepresentation of conflicting evidence can cause harm . |
| Approach: | They propose a method that decomposes texts into atomic statements and uses natural language inference to identify missing facts and a Q A-based metric that extracts question-answer pairs and compares responses across sources. |
| Outcome: | The proposed evaluation metrics show they perform better than more complex metrics, but at a cost. |