INDEX

Explanations

phrases related to something being concealed or not readily apparent

oai_token-act-pair · gpt-3.5-turbo

instances of the word "hidden."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

urses

-0.80

nia

-0.80

SHIP

-0.75

kson

-0.73

annis

-0.71

enegger

-0.70

anwhile

-0.69

night

-0.68

ordan

-0.68

xual

-0.68

POSITIVE LOGITS

 hidden

1.11

 Hidden

1.02

 camoufl

0.87

Hidden

0.86

 behind

0.85

 hatch

0.81

 concealed

0.81

 unseen

0.80

 srfAttach

0.79

 secrets

0.78

Activations Density 0.010%

phrases related to something being concealed or not readily apparent

instances of the word "hidden."

No Comments

No Known Activations