INDEX

Explanations

legal terms and court case related words

oai_token-act-pair · gpt-3.5-turbo

references to legal restrictions or censorship

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xual

-0.69

 Hurricanes

-0.69

 Blessed

-0.66

earned

-0.63

cession

-0.62

signed

-0.61

supp

-0.61

paying

-0.60

guided

-0.60

 Fault

-0.60

POSITIVE LOGITS

gag

1.30

glers

1.22

 reel

0.80

 reflex

0.79

eries

0.78

gey

0.77

weed

0.74

zeb

0.73

bags

0.71

bag

0.71

Activations Density 0.002%

legal terms and court case related words

references to legal restrictions or censorship

No Comments

No Known Activations