INDEX

Explanations

references to the United States

oai_token-act-pair · gpt-3.5-turbo

repeated references to the United States

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

iosyn

-0.80

*/(

-0.80

bably

-0.73

sticks

-0.71

notations

-0.68

»Ĵ

-0.67

charges

-0.66

ovie

-0.66

igs

-0.65

sym

-0.64

POSITIVE LOGITS

 Dollar

0.97

 Embassy

0.97

 Postal

0.89

 Geological

0.89

 Treasury

0.85

 Army

0.83

 embassy

0.81

 Department

0.79

 Government

0.79

 Armed

0.79

Activations Density 0.056%

references to the United States

repeated references to the United States

No Comments

No Known Activations