INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

mble

-0.73

 fused

-0.70

 Tune

-0.69

 eleph

-0.68

 instrument

-0.68

 bonded

-0.67

¶ħ

-0.67

anwhile

-0.67

 captures

-0.66

osate

-0.66

POSITIVE LOGITS

Champ

0.76

ptions

0.73

Benz

0.69

ularity

0.68

esville

0.68

udeb

0.68

ername

0.67

 Plat

0.66

\":

0.66

Reddit

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.