INDEX

Explanations

phrases related to shopping or items being bought or sold

oai_token-act-pair · gpt-3.5-turbo

instances of the word "shop" 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

IBLE

-0.83

ortium

-0.80

sen

-0.71

 Drawing

-0.69

STD

-0.64

 Lieutenant

-0.64

 Maduro

-0.64

Geo

-0.63

 unfocusedRange

-0.63

 Sons

-0.62

POSITIVE LOGITS

sho

1.33

eless

0.94

itsu

0.87

pped

0.85

pper

0.84

isha

0.81

ugen

0.81

pling

0.81

apeshifter

0.81

pload

0.80

Activations Density 0.003%

phrases related to shopping or items being bought or sold

instances of the word "shop" in various contexts

No Comments

No Known Activations