List your company

Grid-world Navigation on FrozenLake reward reversal Hidden drift

100ScoreVanilla + GLOVE
-4235077Jan 27, 2026

Evaluation results

MethodLinks
Vanilla + GLOVEBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.01100
VoyagerBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.01100
Voyager + GLOVEBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.01100
Generative AgentBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.01100
Generative Agent + GLOVEBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.01100
Vanilla +GLOVELLM=Grok-3, Augmentation=GLOVE2026.01100
Voyager +GLOVELLM=Grok-3, Augmentation=GLOVE2026.01100
Voyager + GLOVEBackbone=Llama3.1-8B2026.01100
Vanilla + GLOVELLM Backbone=GPT-4o, Base Agent=Vanilla2026.0197.5
MemoryBank + GLOVELLM Backbone=GPT-4o, Base Agent=MemoryBank2026.0197.5
Voyager + GLOVELLM Backbone=GPT-4o, Base Agent=Voyager2026.0197.5
MemoryBank +GLOVELLM=Grok-3, Augmentation=GLOVE2026.0197.5
Generative Agent +GLOVELLM=Grok-3, Augmentation=GLOVE2026.0197.5
VanillaBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.0195
Vanilla + GLOVEBackbone=Llama3.1-8B2026.0195
Generative Agent + GLOVELLM Backbone=GPT-4o, Base Agent=Generative Agent2026.0192.5
MemoryBankBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.0190
MemoryBank + GLOVEBackbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.0190
VanillaLLM=Grok-32026.0190
Generative Agent + GLOVEBackbone=Llama3.1-8B2026.0190
MemoryBankLLM=Grok-32026.0182.5
MemoryBankBackbone=Llama3.1-8B2026.0165
MemoryBank + GLOVEBackbone=Llama3.1-8B2026.0165
Generative AgentBackbone=Llama3.1-8B2026.0160
MemoryBankLLM Backbone=GPT-4o2026.0157.5
VanillaLLM Backbone=GPT-4o2026.0150
VoyagerLLM Backbone=GPT-4o2026.0150
Generative AgentLLM Backbone=GPT-4o2026.0150
VoyagerLLM=Grok-32026.0150
Generative AgentLLM=Grok-32026.0150
VanillaBackbone=Qwen3-30B, Agent Architecture=Vanilla, GLOVE Augmentation=false2026.0150
Vanilla + GLOVEBackbone=Qwen3-30B, Agent Architecture=Vanilla, GLOVE Augmentation=true2026.0150
VoyagerBackbone=Qwen3-30B, Agent Architecture=Voyager, GLOVE Augmentation=false2026.0150
Voyager + GLOVEBackbone=Qwen3-30B, Agent Architecture=Voyager, GLOVE Augmentation=true2026.0150
Generative AgentBackbone=Qwen3-30B, Agent Architecture=Generative Agent, GLOVE Augmentation=false2026.0150
Generative Agent + GLOVEBackbone=Qwen3-30B, Agent Architecture=Generative Agent, GLOVE Augmentation=true2026.0150
VanillaBackbone=Llama3.1-8B2026.0150
MemoryBank + GLOVEBackbone=Qwen3-30B, Agent Architecture=MemoryBank, GLOVE Augmentation=true2026.0140
MemoryBankBackbone=Qwen3-30B, Agent Architecture=MemoryBank, GLOVE Augmentation=false2026.0137.5
VoyagerBackbone=Llama3.1-8B2026.0135
No Memory (Plain)Backbone=Llama3.3-70B, Drift Type=Implicit, Stage=Hidden drift2026.010
No Memory (Plain)LLM Backbone=GPT-4o2026.010
No Memory (Plain)LLM=Grok-32026.010
No Memory (Plain)Backbone=Qwen3-30B, Agent Architecture=No Memory, GLOVE Augmentation=false2026.010
No Memory (Plain)Backbone=Llama3.1-8B2026.010