INDEX

Explanations

content that has been highlighted or showcased for attention or promotion

oai_token-act-pair · gpt-3.5-turbo

instances of the word "featured" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

etermination

-0.71

 determination

-0.70

prep

-0.67

appro

-0.65

ordinate

-0.64

inking

-0.63

bern

-0.62

mit

-0.62

 calculation

-0.62

wise

-0.62

POSITIVE LOGITS

 featured

3.58

 Featured

2.08

 showcased

1.94

eatured

1.71

Featured

1.58

 feature

1.56

 featuring

1.54

 depicted

1.47

 included

1.39

 features

1.38

Activations Density 0.011%

content that has been highlighted or showcased for attention or promotion

instances of the word "featured" in various contexts

No Comments

No Known Activations