INDEX

Explanations

names starting with 'Ber'

oai_token-act-pair · gpt-3.5-turbo

references to the name "Ber" or similar variations related to individuals or entities associated with that name

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xual

-0.82

uncture

-0.77

eers

-0.75

 Gemini

-0.74

terday

-0.71

Constructed

-0.71

urities

-0.68

 Paran

-0.67

ured

-0.66

okemon

-0.64

POSITIVE LOGITS

rien

1.16

mingham

0.93

rigan

0.93

deen

0.92

Ber

0.92

lus

0.89

keley

0.89

rans

0.85

rett

0.84

wick

0.84

Activations Density 0.007%

names starting with 'Ber'

references to the name "Ber" or similar variations related to individuals or entities associated with that name

No Comments

No Known Activations