INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

wegian

-0.82

Ire

-0.77

uminati

-0.70

afort

-0.68

technical

-0.65

fre

-0.64

asca

-0.64

tera

-0.63

cas

-0.63

well

-0.62

POSITIVE LOGITS

 vest

0.72

ukong

0.70

 sunset

0.67

 Lump

0.66

 Alvarez

0.66

 Pant

0.65

 Garner

0.65

 Jasper

0.65

 Hancock

0.64

prefix

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.