INDEX

Explanations

mentions of the Pope

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "Pope."

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

nesota

-0.82

ript

-0.81

yrinth

-0.79

awks

-0.79

ership

-0.70

sbm

-0.70

ties

-0.68

stract

-0.67

rg

-0.66

awk

-0.66

POSITIVE LOGITS

 Francis

1.15

 Benedict

0.94

 Pope

0.82

 Pablo

0.80

Pope

0.77

 pope

0.74

Father

0.74

 Clement

0.74

otle

0.73

 Patriarch

0.71

Activations Density 0.010%

mentions of the Pope

occurrences of the word "Pope."

No Comments

No Known Activations