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

ents

-0.74

Ð²

-0.70

isks

-0.63

tre

-0.62

Ð¾

-0.62

ÑĮ

-0.60

abeth

-0.60

agg

-0.59

 squared

-0.58

cur

-0.58

POSITIVE LOGITS

 unbeliev

0.64

Oracle

0.64

 eater

0.64

 unden

0.63

krit

0.63

 Brennan

0.63

Ĥİ

0.62

 Ruler

0.62

oxide

0.61

ptin

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.