INDEX

Explanations

words related to belief, disbelief, and faith

oai_token-act-pair · gpt-3.5-turbo

references to belief systems and convictions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

hire

-0.76

apeake

-0.73

amen

-0.70

Mek

-0.70

Interstitial

-0.67

xon

-0.67

assetsadobe

-0.67

 Sieg

-0.64

agher

-0.63

effect

-0.63

POSITIVE LOGITS

ieve

0.98

 belief

0.92

 beliefs

0.90

fulness

0.89

lessness

0.83

 Belief

0.78

tymology

0.77

ieving

0.77

endi

0.76

itious

0.74

Activations Density 0.039%

words related to belief, disbelief, and faith

references to belief systems and convictions

No Comments

No Known Activations