INDEX

Explanations

the abbreviation "WB", possibly related to a specific entity or brand

oai_token-act-pair · gpt-3.5-turbo

references to specific companies or brands

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

xual

-0.80

wagen

-0.78

ateur

-0.77

 delinqu

-0.74

itary

-0.74

yton

-0.73

onyms

-0.73

gas

-0.71

ulic

-0.70

borgh

-0.70

POSITIVE LOGITS

WB

1.32

WB

1.25

KB

0.98

KP

0.95

VB

0.93

FFER

0.89

 Borders

0.89

JD

0.88

KN

0.83

JC

0.82

Activations Density 0.020%

the abbreviation "WB", possibly related to a specific entity or brand

references to specific companies or brands

No Comments

No Known Activations