INDEX

Explanations

phrases related to offering a perspective or opinion on a topic

oai_token-act-pair · gpt-3.5-turbo

references to different perspectives or interpretations (takes) on various subjects

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

 constitu

-0.77

ĺħ

-0.75

 Aires

-0.69

 Buildings

-0.68

 certific

-0.67

 obser

-0.65

eger

-0.65

 Places

-0.65

 Standards

-0.63

 thous

-0.62

POSITIVE LOGITS

away

1.29

aways

1.14

over

1.05

athon

1.03

able

1.02

back

1.00

backs

0.99

ings

0.97

outs

0.93

ability

0.92

Activations Density 0.272%

phrases related to offering a perspective or opinion on a topic

references to different perspectives or interpretations (takes) on various subjects

No Comments

No Known Activations