INDEX

Explanations

items within a range of values or categories

oai_token-act-pair · gpt-3.5-turbo

phrases that describe a range or variety of items or conditions

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

Drive

-0.65

illard

-0.63

ocard

-0.62

amus

-0.58

 Official

-0.58

 thing

-0.57

athan

-0.57

gin

-0.56

daq

-0.56

vich

-0.56

POSITIVE LOGITS

 ranging

3.46

ranging

1.99

 ranged

1.76

 varying

1.54

 ranges

1.53

 spanning

1.51

 totaling

1.47

 range

1.37

 including

1.23

 averaging

1.16

Activations Density 0.017%

items within a range of values or categories

phrases that describe a range or variety of items or conditions

No Comments

No Known Activations