INDEX

Explanations

words related to religious figures, such as "nun," "Sister," and "convent."

oai_token-act-pair · gpt-3.5-turbo

references to nuns and sisterhood

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

à¸

-0.72

resp

-0.69

ullivan

-0.67

eters

-0.67

Lex

-0.64

iku

-0.64

asper

-0.63

izophren

-0.62

acqu

-0.62

OTO

-0.62

POSITIVE LOGITS

nery

1.06

nun

1.05

 nuns

1.02

 Sisters

0.90

 Sister

0.85

 Bride

0.83

 convent

0.80

 choir

0.78

nai

0.76

ella

0.74

Activations Density 0.015%

words related to religious figures, such as "nun," "Sister," and "convent."

references to nuns and sisterhood

No Comments

No Known Activations