INDEX

Explanations

references to hiding or being hidden

oai_token-act-pair · gpt-3.5-turbo

phrases related to concealment or hidden activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Pwr

-0.77

apest

-0.75

lez

-0.75

 roundup

-0.68

equal

-0.66

trak

-0.65

 Lowell

-0.62

Grade

-0.61

spir

-0.60

 à¤

-0.60

POSITIVE LOGITS

 secrets

0.87

 concealed

0.79

anwhile

0.78

ĸļ

0.77

 disgu

0.74

 secret

0.74

 obscured

0.73

hid

0.73

 hidden

0.73

 undet

0.73

Activations Density 0.260%

references to hiding or being hidden

phrases related to concealment or hidden activities

No Comments

No Known Activations