INDEX

Explanations

phrases related to scandals or controversies

oai_token-act-pair · gpt-3.5-turbo

variations of the word "gate."

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

arya

-0.72

ľ

-0.65

aldi

-0.64

accompan

-0.64

@#&

-0.63

ynt

-0.63

 subp

-0.62

podcast

-0.62

inances

-0.62

Ĥİ

-0.62

POSITIVE LOGITS

gate

1.13

way

0.97

Gate

0.90

ways

0.90

boro

0.88

keeper

0.79

plot

0.75

ardless

0.75

door

0.75

ble

0.74

Activations Density 0.007%

phrases related to scandals or controversies

variations of the word "gate."

No Comments

No Known Activations