INDEX

Explanations

the word "cleavage"

oai_token-act-pair · gpt-3.5-turbo

words starting with "cle" that are potentially related to clarity or cleanliness

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

Not in Any Lists

No Comments

Negative Logits

bucks

-0.77

ãĥ¼ãĥĨãĤ£

-0.64

NPR

-0.60

wagen

-0.59

 Galaxy

-0.59

 Beckham

-0.59

 sqor

-0.58

 counting

-0.58

 Slug

-0.57

 representing

-0.57

POSITIVE LOGITS

veland

1.30

avage

1.28

arest

1.23

vere

1.11

aved

1.10

aving

1.10

aves

1.03

combe

1.00

ave

0.96

vern

0.94

Activations Density 0.014%

the word "cleavage"

words starting with "cle" that are potentially related to clarity or cleanliness

No Comments

No Known Activations