INDEX

Explanations

mentions of a specific entity likely named 'Snopes'

oai_token-act-pair · gpt-3.5-turbo

references to a specific character or entity identified as "Sn."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 secrecy

-0.77

xual

-0.75

 Sunni

-0.71

TAIN

-0.68

 Hezbollah

-0.68

heid

-0.67

 patriarchy

-0.64

 Sikh

-0.64

 PRESS

-0.63

 directions

-0.62

POSITIVE LOGITS

icket

1.10

agging

1.10

ipers

1.06

appy

1.06

akes

1.06

agged

1.03

ook

1.01

oot

1.01

ugg

1.00

ipes

0.99

Activations Density 0.009%

mentions of a specific entity likely named 'Snopes'

references to a specific character or entity identified as "Sn."

No Comments

No Known Activations