INDEX

Explanations

phrases related to a strong emotional reaction or impact, especially those involving surprise or disturbance

oai_token-act-pair · gpt-3.5-turbo

words related to shock or surprising events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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 10 - surprise

No Comments

Negative Logits

ickr

-0.72

 Haku

-0.66

onymous

-0.64

etheless

-0.64

ublic

-0.63

uties

-0.61

 Princ

-0.61

SHIP

-0.60

allery

-0.59

tein

-0.59

POSITIVE LOGITS

waves

1.42

wave

1.40

ingly

1.21

 absor

1.12

tro

0.92

er

0.92

crow

0.86

fully

0.79

mong

0.77

bang

0.75

Activations Density 0.033%

phrases related to a strong emotional reaction or impact, especially those involving surprise or disturbance

words related to shock or surprising events

No Comments

No Known Activations