INDEX

Explanations

sequences indicating a continuation or completion of actions

oai_token-act-pair · gpt-3.5-turbo

transitional phrases or sequences in narratives

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 IMAGES

-0.63

ense

-0.61

sid

-0.61

natureconservancy

-0.60

feature

-0.59

smart

-0.57

ourage

-0.56

hea

-0.54

é¾įå¥ĳå£«

-0.54

 portrayal

-0.53

POSITIVE LOGITS

 finally

1.22

 Finally

1.19

 again

1.10

Finally

1.06

 thence

1.03

 Again

1.01

Then

0.97

 thereafter

0.96

again

0.95

 then

0.94

Activations Density 0.602%

sequences indicating a continuation or completion of actions

transitional phrases or sequences in narratives

No Comments

No Known Activations