INDEX

Explanations

words related to personal beliefs and religion

oai_token-act-pair · gpt-3.5-turbo

mentions of beliefs, particularly in the context of religion and personal values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Mamm

-0.80

agher

-0.77

nie

-0.70

amen

-0.68

 Amen

-0.68

Mek

-0.67

shows

-0.66

xon

-0.66

assetsadobe

-0.66

GV

-0.65

POSITIVE LOGITS

 beliefs

1.13

 belief

0.92

ieve

0.85

 Belief

0.85

 mant

0.82

 tenets

0.79

linger

0.79

ophical

0.79

kie

0.77

cius

0.76

Activations Density 0.037%

words related to personal beliefs and religion

mentions of beliefs, particularly in the context of religion and personal values

No Comments

No Known Activations