INDEX

Explanations

phrases related to specific positions or states

oai_token-act-pair · gpt-3.5-turbo

phrases related to being in a particular position or situation

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

 Flavoring

-0.97

cause

-0.75

rosso

-0.72

Interstitial

-0.72

glomer

-0.71

favorite

-0.69

âĵĺ

-0.68

etti

-0.66

 Medals

-0.65

 Kard

-0.62

POSITIVE LOGITS

 position

0.73

ocene

0.72

pora

0.71

utherland

0.69

porting

0.69

holders

0.69

vati

0.67

ioned

0.67

ession

0.67

holder

0.66

Activations Density 0.026%

phrases related to specific positions or states

phrases related to being in a particular position or situation

No Comments

No Known Activations