INDEX

Explanations

for threats or words related to threats, such as genocide, theft, and endangerment

oai_token-act-pair · gpt-3.5-turbo

terms related to threats or dangers

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

Downloadha

-0.96

Interstitial

-0.70

 Sunrise

-0.68

 Idle

-0.67

cules

-0.67

VILLE

-0.67

perty

-0.64

 Coul

-0.64

 Scotia

-0.63

Spoiler

-0.63

POSITIVE LOGITS

ened

1.70

ening

1.55

eners

1.22

ener

1.19

en

1.18

ens

1.06

aced

0.98

eny

0.94

ethy

0.90

nant

0.85

Activations Density 0.022%

for threats or words related to threats, such as genocide, theft, and endangerment

terms related to threats or dangers

No Comments

No Known Activations