INDEX

Explanations

the word "studio"

oai_token-act-pair · gpt-3.5-turbo

references to studios, particularly for music or film recording

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

pend

-0.71

avez

-0.66

isms

-0.65

abad

-0.63

 refere

-0.62

vati

-0.62

bh

-0.61

reads

-0.60

theless

-0.59

 constitu

-0.59

POSITIVE LOGITS

 studio

0.95

 studios

0.85

orpor

0.78

ctor

0.76

rats

0.73

smanship

0.73

camp

0.72

mog

0.69

rador

0.69

venant

0.67

Activations Density 0.027%

the word "studio"

references to studios, particularly for music or film recording

No Comments

No Known Activations