INDEX

Explanations

references to the United States of America (USA)

oai_token-act-pair · gpt-3.5-turbo

mentions of the United States (USA)

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

rained

-0.84

flush

-0.80

 garg

-0.73

rums

-0.72

 bath

-0.71

lords

-0.71

dress

-0.67

 quarters

-0.67

lers

-0.66

lines

-0.64

POSITIVE LOGITS

BIL

1.04

USA

0.97

BILITY

0.96

terday

0.93

ADA

0.88

EGA

0.87

icago

0.84

ITE

0.82

xit

0.81

 TODAY

0.80

Activations Density 0.004%

references to the United States of America (USA)

mentions of the United States (USA)

No Comments

No Known Activations