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

 Ukrainians

-0.81

Ily

-0.78

 fears

-0.74

 Pavel

-0.70

 worries

-0.69

 Emin

-0.69

 Yuri

-0.67

Dru

-0.66

Aly

-0.66

 Russians

-0.65

POSITIVE LOGITS

à©

0.90

ãĥīãĥ©

0.79

orest

0.71

à¨

0.71

Ü

0.70

Ö¼

0.70

ocal

0.69

ogl

0.68

ãĥ¼ãĥĨ

0.68

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.