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

airs

-0.71

ugs

-0.68

arted

-0.68

Ago

-0.67

ocular

-0.67

urations

-0.66

tee

-0.66

Eye

-0.65

ãĤ´ãĥ³

-0.65

uming

-0.64

POSITIVE LOGITS

okia

0.68

 guiActiveUn

0.63

onomous

0.61

NCT

0.61

mor

0.60

 Town

0.60

 Sask

0.59

ovo

0.59

 Zamb

0.58

 Feder

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.