INDEX

Explanations

phrases related to official or unofficial authority and leadership roles

oai_token-act-pair · gpt-3.5-turbo

phrases indicating "de facto" concepts or statuses

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

ander

-0.74

 Symphony

-0.72

 Braz

-0.71

iments

-0.70

 Hungry

-0.69

elo

-0.69

mong

-0.69

etter

-0.68

ulz

-0.68

lyss

-0.68

POSITIVE LOGITS

 facto

0.84

ignment

0.83

 monarchy

0.78

dom

0.75

 guiIcon

0.70

 guiActiveUn

0.68

 adherence

0.67

 compliance

0.67

 stabil

0.66

TY

0.66

Activations Density 0.019%

phrases related to official or unofficial authority and leadership roles

phrases indicating "de facto" concepts or statuses

No Comments

No Known Activations