INDEX

Explanations

instances of the word "custody" in various contexts

oai_token-act-pair · gpt-3.5-turbo

references to individuals being in custody

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

gnu

-0.76

olf

-0.72

hof

-0.71

gom

-0.68

audi

-0.66

Ign

-0.65

Downloadha

-0.63

die

-0.62

hist

-0.61

 hypocritical

-0.60

POSITIVE LOGITS

 custody

1.19

 Cust

1.02

holding

0.82

agy

0.78

tesy

0.76

 custod

0.76

taker

0.75

chwitz

0.73

 Clerk

0.73

estation

0.71

Activations Density 0.010%

instances of the word "custody" in various contexts

references to individuals being in custody

No Comments

No Known Activations