INDEX

Explanations

phrases related to advocating or supporting a particular cause or action

oai_token-act-pair · gpt-3.5-turbo

instances of the word "advocate" and its variants

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bourne

-0.69

Lo

-0.68

brance

-0.68

Sus

-0.68

omaly

-0.66

hook

-0.66

TM

-0.63

few

-0.62

kes

-0.62

six

-0.62

POSITIVE LOGITS

 passionately

0.92

 advocating

0.87

 vigorously

0.79

 stances

0.79

 vehemently

0.78

 moderation

0.77

 advocated

0.75

 tirelessly

0.75

 compromises

0.74

 anew

0.73

Activations Density 0.076%

phrases related to advocating or supporting a particular cause or action

instances of the word "advocate" and its variants

No Comments

No Known Activations