SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2405.05348 · ICLR · 2024

The Effect of Model Size on LLM Post-hoc Explainability via LIME

Noah Siegel, Amy Widdicombe, Henning Heyen, María Pérez-Ortiz, Philip Treleaven

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 3 functions out of this paper's own repositories and ran 1 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
henningheyen/scalability-of-llm-posthoc-explanations canonical 1 of 3
FunctionStatusWhere it lives
make_test_set_esnli Ran henningheyen/scalability-of-llm-posthoc-explanations/utils.py
code served (permissive licence) · get_code("e1c087e8e537738f")
make_test_set_cose Not yet run henningheyen/scalability-of-llm-posthoc-explanations/utils.py
code served (permissive licence) · get_code("2fbae2b2de24931f")
make_test_set_mnli Not yet run henningheyen/scalability-of-llm-posthoc-explanations/utils.py
code served (permissive licence) · get_code("d428e5283a97296e")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Large language models (LLMs) are becoming bigger to boost performance. However, little is known about how explainability is affected by this trend. This work explores LIME explanations for DeBERTaV3 models of four different sizes on natural language inference (NLI) and zero-shot classification (ZSC) tasks. We evaluate the explanations based on their faithfulness to the models' internal decision processes and their plausibility, i.e. their agreement with human explanations. The key finding is that increased model size does not correlate with plausibility despite improved model performance, suggesting a misalignment between the LIME explanations and the models' internal processes as model size increases. Our results further suggest limitations regarding faithfulness metrics in NLI contexts.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2405.05348")
get_code_for_paper("2405.05348")
have("2405.05348")

Connect an agent — have() is free.