INDEX

Explanations

references to "idol" or "idolizing" in a text

oai_token-act-pair · gpt-3.5-turbo

terms related to idols or idolization in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

RAW

-0.72

-0.70

aunder

-0.69

arnaev

-0.66

~~~~~~~~~~~~~~~~

-0.66

--------------------------------------------------------

-0.66

Klu

-0.65

utenberg

-0.64

cale

-0.63

 unintention

-0.63

POSITIVE LOGITS

 idol

1.16

 worshipped

0.93

 idols

0.92

 Idol

0.88

ãħĭãħĭ

0.88

ãħĭ

0.85

 worsh

0.83

 worship

0.81

 sacrific

0.81

imates

0.81

Activations Density 0.008%

references to "idol" or "idolizing" in a text

terms related to idols or idolization in various contexts

No Comments

No Known Activations