When Not to Trust Language Models: Investigating Effectiveness and Limitations of Parametric and Non-Parametric Memories

By Alex Troy Mallen, Akari Asai, Victor Zhong, Rajarshi Das, Hannaneh Hajishirzi, Daniel Khashabi. 2022

199 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Management Science and Operations Research, Large Language Models, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Management Science and Operations Research

In Large Language Models

In Natural Language Processing

Citations in the corpus

Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling (2023, background): “Recent work has identified the frequency of specific facts within a corpus as an important factor in how likely a model is capable of applying that fact in response to a natural language question (Razeghi et al., 2022; Elazar et al., 2022; Kandpal et al., 2022; Mallen et al., 2022).”
ToolQA: A Dataset for LLM Question Answering with External Tools (2023, background): “Concurrent to our work, there are also several studies [70, 37] that investigate the parametric and nonparametric knowledge from LLMs.”
Unifying Large Language Models and Knowledge Graphs: A Roadmap (2023, background): “Alex et al. [130] investigate the capacity of LLMs to retain less popular factual knowledge.”
A Comprehensive Evaluation of Tool-Assisted Generation Strategies (2023, background): “It has been shown that LMs have difficulty solving problems involving long-tail entities (Kandpal et al., 2022; Mallen et al., 2022) and complex mathematical reasoning challenges (Mishra et al.”
“According to . . . ”: Prompting Language Models Improves Quoting from Pre-Training Data (2023, methodology): “…to alleviate this problem in various ways, including retrieving grounded documents before generation (Sun et al., 2023; Borgeaud et al., 2022; Mallen et al., 2023; Weller et al., 2022), applying new decoding approaches (He et al., 2022), post hoc tuning of LLMs (Menick et al., 2022; Lee et…”
Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learning (2025, unknown): “We evaluate S EARCH -R1 on seven benchmark datasets, categorized as follows: (1) General Question Answering : NQ (Kwiatkowski et al., 2019), TriviaQA (Joshi et al., 2017), and PopQA (Mallen et al., 2022).”
Adaptive Chameleon or Stubborn Sloth: Revealing the Behavior of Large Language Models in Knowledge Conflicts (2023, methodology): “Such a framework, which has proven its efficacy in enhancing large language models (Shi et al., 2023b; Ram et al., 2023; Mallen et al., 2022), is adopted in real-world applications such as New Bing and ChatGPT Plugins.”
TÜLU 3: Pushing Frontiers in Open Language Model Post-Training (2024, unknown): “PopQA [Mallen et al., 2022] is an entity-centric question-answering benchmark that evaluates language models’ tendency to forget information about long-tail entities.”
Retrieval-based Language Models and Applications (2023, background): “Retrieval-based LMs can outperform LMs without retrieval by a large margin with much fewer parameters (Mallen et al., 2022), can update their knowledge by replacing their retrieval corpora (Izacard et al., 2022b), and provide citations for users to easily verify and evaluate the predictions (Menick…”
REPLUG: Retrieval-Augmented Black-Box Language Models (2023, methodology): “While concurrent work (Mallen et al., 2022; Si et al., 2023; Yu et al., 2023; Khattab et al., 2022) has demonstrated that using a frozen retriever can improve GPT-3 performance on open-domain question answering, we approach the problem in a more general setting, including language modeling and…”
FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation (2023, background): “, 2021), or focuses on question answering with short answers (Kadavath et al., 2022; Kandpal et al., 2022; Mallen et al., 2022; Nori et al., 2023).”
Active Retrieval Augmented Generation (2023, background): “Adaptive retrieval has been studied in single-time retrieval scenarios based on either question popularity or generation probabilities (Mallen et al., 2022; Li et al., 2023), while we focus on long-form generation requiring active information access.”
Knowledge Conflicts for LLMs: A Survey (2024, methodology): “To enhance robustness, Hong et al. (2023) propose a fine-tuning method that trains a discriminator and decoder simultaneously using a shared encoder, alongside strategies involving prompting GPT-3 to identify perturbed documents and integrating the discriminator’s output into prompts.”