INDEX

Explanations

mentions of different states in a geopolitical context

oai_token-act-pair · gpt-3.5-turbo

references to geographical locations, particularly states and regions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Horowitz

-0.99

Pie

-0.87

 Lennon

-0.84

Zup

-0.81

 Hoffman

-0.80

Leo

-0.80

 Dough

-0.79

 Ogre

-0.78

 Berger

-0.77

 Pizza

-0.76

POSITIVE LOGITS

STATE

1.76

State

1.72

state

1.71

States

1.68

 State

1.66

 STATE

1.64

 state

1.63

 states

1.57

 States

1.54

states

1.48

Activations Density 0.266%

mentions of different states in a geopolitical context

references to geographical locations, particularly states and regions

No Comments

No Known Activations