INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

pex

-0.72

 punishable

-0.66

ribune

-0.65

HT

-0.64

 priced

-0.64

oria

-0.63

 rated

-0.63

 redistributed

-0.62

 trimmed

-0.61

stadt

-0.59

POSITIVE LOGITS

nsic

0.71

natureconservancy

0.67

aughs

0.66

Kin

0.65

Maria

0.64

general

0.64

Unity

0.64

acqu

0.63

vas

0.63

uve

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.