AI Paper Atlas

Research map / Large language models / Agents, code and math

Language Models Meet World Models: Embodied Experiences Enhance Language Models

Jiannan Xiang, Tianhua Tao, Yi Gu, Tianmin Shu, Zirui Wang, Zichao Yang, Zhiting Hu

NeurIPS 2023 · Large language models › Agents, code and math · approach: Long-horizon LLM agents · 23 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
finetune on embodied experiences from VirtualHome with EWC-LoRAGPT-J-6B · Object Path TrackingLCS: 33.86 -> 98.67Improved
finetune with E2WM (EWC-LoRA)GPT-J-6B · Countingaccuracy: 30.41 -> 67.01Improved
finetune with E2WM (EWC-LoRA)GPT-J-6B · Plan GenRouge-L: 34.31 -> 51.23Improved
finetune with E2WM (EWC-LoRA)GPT-J-6B · Pile test subsetperplexity: 3.443 -> 3.537No clear effect
add EWC regularization to LoRA (ablation)LoRA finetuning · Pile test subset (GPT-Neo)perplexity: 4.360 -> 4.193Improved
use LoRA instead of full-parameter EWC (ablation)EWC · Pile test subset (GPT-Neo)perplexity: 4.995 -> 4.360Improved
replace EWC with KL penalty (ablation)LoRA finetuning · Pile test subset (GPT-Neo)perplexity: 4.193 -> 5.029Hurt
add EWC to LoRA (ablation)LoRA finetuning · Counting (GPT-Neo)accuracy: 27.84 -> 28.87No clear effect
add plan generation training task (ablation)E2WM GPT-Neo · Plan GenRouge-L: 14.48 -> 49.70Improved
add activity recognition training task (ablation)E2WM GPT-Neo · Act Recogaccuracy: 48.63 -> 85.43Improved
add object path tracking training task (ablation)E2WM GPT-Neo · Obj PTscore: 29.90 -> 85.91Improved
add counting training task (ablation)E2WM GPT-Neo · Countingaccuracy: 35.05 -> 28.87Hurt

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