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

UTERS

-0.83

¼

-0.78

¦

-0.75

ãĥĩãĤ£

-0.73

ESCO

-0.70

tained

-0.69

riad

-0.69

buster

-0.68

wcsstore

-0.66

ati

-0.66

POSITIVE LOGITS

 Philly

0.72

uyomi

0.72

 Surrey

0.69

 Charleston

0.67

 Nottingham

0.67

 Execution

0.66

 Marse

0.64

 damned

0.64

 marsh

0.62

illin

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.