INDEX

Explanations

keywords related to news headlines such as events, institutions, people, and places

oai_token-act-pair · gpt-3.5-turbo

instances of the word "CLOSE" in various contexts

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

hid

-0.91

ologically

-0.89

audi

-0.77

ozo

-0.73

uana

-0.73

 alienated

-0.72

urb

-0.72

gery

-0.72

informed

-0.69

unal

-0.67

POSITIVE LOGITS

CLOSE

1.22

UTERS

0.83

GROUND

0.80

 ALSO

0.77

 SPACE

0.76

 Close

0.75

aign

0.75

 PHOTO

0.75

 STORY

0.75

 Remastered

0.74

Activations Density 0.014%

keywords related to news headlines such as events, institutions, people, and places

instances of the word "CLOSE" in various contexts

No Comments

No Known Activations