INDEX

Explanations

the word "Cedar"

oai_token-act-pair · gpt-3.5-turbo

references to places or organizations associated with "Cedar."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

enegger

-1.13

arty

-0.91

ptin

-0.80

andowski

-0.80

imester

-0.79

ntil

-0.78

psons

-0.77

merce

-0.77

ierrez

-0.77

oulos

-0.76

POSITIVE LOGITS

 Cedar

1.18

 Rapids

1.01

edar

0.94

 Ruler

0.86

Winged

0.85

mere

0.80

Constructed

0.79

Saint

0.78

rate

0.76

Series

0.76

Activations Density 0.016%

the word "Cedar"

references to places or organizations associated with "Cedar."

No Comments

No Known Activations