INDEX

Explanations

names of high schools

oai_token-act-pair · gpt-3.5-turbo

references to high schools

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

 rall

-0.81

vier

-0.81

oidal

-0.78

ĸļ

-0.77

å¥³

-0.73

arium

-0.72

pus

-0.71

Downloadha

-0.70

afort

-0.70

ä»£

-0.68

POSITIVE LOGITS

landers

1.16

lighting

0.99

nesses

0.92

lander

0.90

lights

0.90

 School

0.87

rises

0.87

school

0.87

ness

0.86

land

0.81

Activations Density 0.030%

names of high schools

references to high schools

No Comments

No Known Activations