Papers by Amos Azaria
Multi-Relational Question Answering from Narratives: Machine Reading and Reasoning in Simulated Worlds (P18-1)
Copied to clipboard
| Challenge: | Question Answering (QA) has primarily focused on knowledge bases or free text as a source of knowledge. |
| Approach: | They propose a task of multi-relational QA over personal narrative using text worlds . they generate and release a lightweight Python-based framework for easily generating additional worlds and narrative . |
| Outcome: | The proposed framework combines elements of structured QA over knowledge bases and unstructured QA . it generates and analyzes five diverse datasets with dynamic narrative . the framework is lightweight and easy to use . |
The Internal State of an LLM Knows When It’s Lying (2023.findings-emnlp)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have shown exceptional performance in various tasks, but one of their main drawbacks is generating inaccurate or false information with a confident tone. |
| Approach: | They propose to train a classifier that outputs the probability that a statement is truthful based on the hidden layer activations of the LLM as it reads or generates the statement. |
| Outcome: | The proposed classifier achieves an average of 71% to 83% accuracy labeling which sentences are true versus false, depending on the LLM base model. |