INDEX

Explanations

phrases related to drive-ins or drive-throughs

oai_token-act-pair · gpt-3.5-turbo

phrases related to drive-through services

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

 Arabian

-0.68

Vel

-0.66

ciating

-0.64

 Monsters

-0.63

 Vaughn

-0.62

 Dimension

-0.60

GHC

-0.59

Lum

-0.59

 Chili

-0.58

 Cycl

-0.58

POSITIVE LOGITS

through

1.33

off

1.27

offs

1.25

out

1.20

bys

1.17

around

1.15

outs

1.15

forward

1.13

ahead

1.12

away

1.10

Activations Density 0.058%

phrases related to drive-ins or drive-throughs

phrases related to drive-through services

No Comments

No Known Activations