AI Paper Atlas

Research map / Large language models / Retrieval and knowledge

Statistical Knowledge Assessment for Large Language Models

Qingxiu Dong, Jingjing Xu, Lingpeng Kong, Zhifang Sui, Lei Li

NeurIPS 2023 · Large language models › Retrieval and knowledge · approach: Knowledge-augmented question answering · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
instruction-tune LLaMA-13B on Alpaca dataLLaMA-13B · KaRR T-RExKaRR score: 13.86 -> 8.24Hurt
instruction-tune LLaMA-13B on Vicuna data from a stronger modelLLaMA-13B · KaRR T-RExKaRR score: 13.86 -> 19.50Improved
instruction-tune LLaMA-13B on WizardLM dataLLaMA-13B · KaRR T-RExKaRR score: 13.86 -> 16.90Improved
scale LLaMA from 7B to 13B to 65BLLaMA · KaRR T-RExKaRR score: 12.37 -> 13.86 -> 14.56Improved
replace T5-3B with GPT-NEO of similar sizepretrained small LMs · KaRR T-RExKaRR score: 9.52 -> 13.44Improved
average probability over k paraphrased prompts (K-Prompts) instead of latent-variable ratio (ablation)KaRR · Human eval GPT2-XLKendall tau: 0.43 -> 0.32Hurt
use KaRR instead of LAMA@1 for detecting unknown factsknowledge assessment · Human eval GPT2-XLrecall: 83.25% -> 95.18%Improved
use KaRR instead of K-Prompts under paraphrased templates (ablation)knowledge assessment · Paraphrased templatesvariance: 2.34 -> 0.67Improved
use KaRR instead of LAMA@10 on false synthetic factsknowledge assessment · Spurious correlationSP: 64.29 -> 1.94Improved
use KaRR instead of K-Prompts on spurious positivesknowledge assessment · Spurious correlationSP: 0.00 vs 1.94No clear effect
increase sampling number K (ablation)KaRR GPT2-XL · Variance and human correlationvariance and correlationImproved
scale OPT from 350M to 2.7B to 175BOPT · Best/worst-known relationsKaRR scoreImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy