INDEX

Explanations

years from the 1600s and 1800s

oai_token-act-pair · gpt-3.5-turbo

references to specific historical centuries

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bid

-1.00

unal

-0.95

avering

-0.94

abol

-0.94

alid

-0.88

bor

-0.88

uana

-0.85

abor

-0.83

ork

-0.80

andering

-0.80

POSITIVE LOGITS

mAh

1.02

 certs

0.92

MHz

0.91

0.85

MHz

0.82

0.81

0.80

0.79

ãĤ¦ãĤ¹

0.77

Activations Density 0.009%

years from the 1600s and 1800s

references to specific historical centuries

No Comments

No Known Activations