INDEX

Explanations

the word "Show" with varying activation strengths

oai_token-act-pair · gpt-3.5-turbo

instances of the word "Show"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

OAD

-0.72

ADE

-0.67

adem

-0.61

bsite

-0.59

 Seeking

-0.59

 Bere

-0.58

ngth

-0.58

auri

-0.57

eco

-0.57

 Âł

-0.56

POSITIVE LOGITS

 Thumbnails

1.03

alter

0.74

cases

0.70

case

0.68

me

0.64

 nested

0.63

boat

0.63

kat

0.61

downs

0.60

boats

0.59

Activations Density 0.028%

the word "Show" with varying activation strengths

instances of the word "Show"

No Comments

No Known Activations