INDEX

Explanations

quantitative information or probabilistic statements

oai_token-act-pair · gpt-3.5-turbo

phrases related to statistical probabilities and numerical evaluation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

Trend

-0.68

cour

-0.65

entimes

-0.65

vertising

-0.63

ÃĥÃĤ

-0.62

focused

-0.62

udos

-0.62

Lots

-0.62

fortunately

-0.62

mmm

-0.61

POSITIVE LOGITS

 Auschwitz

1.00

SCP

0.97

 Voldemort

0.94

 Canaver

0.90

 comet

0.89

 Tsarnaev

0.87

WTC

0.86

 embryos

0.84

 transsexual

0.84

 Bigfoot

0.83

Activations Density 1.020%

quantitative information or probabilistic statements

phrases related to statistical probabilities and numerical evaluation

No Comments

No Known Activations