INDEX

Explanations

phrases indicating inclusivity or emphasis

oai_token-act-pair · gpt-3.5-turbo

emphatic expressions and inclusive language

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ipel

-0.74

adem

-0.74

ord

-0.71

opol

-0.71

odore

-0.70

ãĥĥãĥĪ

-0.66

lahoma

-0.66

perties

-0.66

roman

-0.66

alk

-0.65

POSITIVE LOGITS

THING

1.24

 THERE

1.20

ELY

1.17

 FANTASY

1.15

 FUCK

1.14

 MUCH

1.10

INESS

1.09

VER

1.08

FORE

1.08

FORM

1.08

Activations Density 0.129%

phrases indicating inclusivity or emphasis

emphatic expressions and inclusive language

No Comments

No Known Activations