INDEX

Explanations

mentions of American entities or individuals

oai_token-act-pair · gpt-3.5-turbo

references to American entities or culture

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

zik

-0.89

heed

-0.83

orders

-0.75

ende

-0.74

cffffcc

-0.74

ijk

-0.73

illet

-0.72

bley

-0.72

stretched

-0.68

hod

-0.68

POSITIVE LOGITS

 Airlines

0.99

ICAN

0.98

 Idol

0.91

 Samoa

0.87

 Legion

0.87

 Express

0.85

ican

0.80

 Horror

0.75

 Indian

0.74

 Lutheran

0.73

Activations Density 0.046%

mentions of American entities or individuals

references to American entities or culture

No Comments

No Known Activations