INDEX

Explanations

No Explanations Found

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

 nerve

-0.71

 Orion

-0.63

 mankind

-0.62

 humanity

-0.61

Interstitial

-0.60

 sights

-0.60

 Garcia

-0.59

 Rome

-0.58

 nerves

-0.57

 rated

-0.56

POSITIVE LOGITS

$$$$

0.87

Bye

0.79

olicy

0.77

yip

0.76

senal

0.76

leans

0.76

afety

0.71

schild

0.70

conservancy

0.69

elsius

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.