INDEX

Explanations

negations, indicating what something is not

oai_token-act-pair · gpt-3.5-turbo

instances where something is stated as not occurring or being present

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

LG

-0.75

ngth

-0.68

ixel

-0.68

creen

-0.68

 Tours

-0.65

¿½

-0.64

kamp

-0.64

hower

-0.64

gha

-0.63

ĸļ

-0.62

POSITIVE LOGITS

 necessarily

1.27

icably

1.16

icable

1.13

epad

1.09

withstanding

1.06

etheless

1.05

eworthy

0.96

 bothered

0.91

 bothering

0.88

ifies

0.85

Activations Density 0.065%

negations, indicating what something is not

instances where something is stated as not occurring or being present

No Comments

No Known Activations