List your company

State Unlearning on AlfWorld

100Unlearn EfficacyNL (GPT)
9597.5100102.5Apr 1, 2026

Evaluation results

MethodLinks
NL (GPT)Prompting method=NL, Backbone LLM=GPT2026.04100989593
NL (Claude)Prompting method=NL, Backbone LLM=Claude2026.04100858382
NL (Llama)Prompting method=NL, Backbone LLM=Llama2026.04100939493
Code (GPT)Prompting method=Code, Backbone LLM=GPT2026.04100949589
Code (Claude)Prompting method=Code, Backbone LLM=Claude2026.04100798585
Code (Llama)Prompting method=Code, Backbone LLM=Llama2026.04100899079
Exam. (GPT)Prompting method=Example-based, Backbone LLM=GPT2026.04100962013
Exam. (Claude)Prompting method=Example-based, Backbone LLM=Claude2026.041007978
Exam. (Llama)Prompting method=Example-based, Backbone LLM=Llama2026.04100854044