INDEX

Explanations

exclamatory questions expressing surprise or disbelief

oai_token-act-pair · gpt-3.5-turbo

phrases that express confusion or surprise, often punctuated with exclamatory or interrogative marks

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

•layer 1 - surprise

No Comments

Negative Logits

Ö¼

-0.78

amen

-0.73

hob

-0.71

itions

-0.70

odor

-0.68

 corrid

-0.68

 purified

-0.68

missions

-0.68

rive

-0.68

iral

-0.67

POSITIVE LOGITS

?!

1.52

?,

1.03

!?

1.02

?!"

1.01

!!

0.99

#$

0.98

!!!!!

0.98

Huh

0.98

??

0.97

 Anyway

0.96

Activations Density 0.008%

exclamatory questions expressing surprise or disbelief

phrases that express confusion or surprise, often punctuated with exclamatory or interrogative marks

No Comments

No Known Activations