INDEX

Explanations

information about where people live

oai_token-act-pair · gpt-3.5-turbo

instances of the word "lives" in reference to individuals' residences

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

sonian

-0.81

elight

-0.76

enei

-0.72

ion

-0.71

ery

-0.71

anche

-0.69

essee

-0.68

xual

-0.67

oug

-0.67

chy

-0.65

POSITIVE LOGITS

lihood

0.77

 Wage

0.77

 upstairs

0.75

behind

0.74

blog

0.73

chool

0.72

 Rent

0.70

vic

0.70

 Juliet

0.69

journal

0.69

Activations Density 0.032%

information about where people live

instances of the word "lives" in reference to individuals' residences

No Comments

No Known Activations