INDEX

Explanations

words related to emphasis, certainty, and negation

oai_token-act-pair · gpt-3.5-turbo

strongly negative or absolute expressions

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

odore

-0.75

ipel

-0.74

inav

-0.71

igration

-0.71

opol

-0.68

adem

-0.66

iasis

-0.65

urers

-0.64

midt

-0.64

lahoma

-0.63

POSITIVE LOGITS

THING

1.20

 THERE

1.20

ELY

1.19

 MUCH

1.10

 FUCK

1.07

VER

1.07

 THEN

1.03

NOT

1.01

ADA

1.01

BE

1.01

Activations Density 0.115%

words related to emphasis, certainty, and negation

strongly negative or absolute expressions

No Comments

No Known Activations