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

 encount

-0.95

tesy

-0.90

taboola

-0.87

Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯Â¯

-0.85

BIL

-0.76

ĳå£«

-0.75

 proble

-0.74

 "$:/

-0.74

 ï¿½ï¿½ï¿½ï¿½

-0.73

natureconservancy

-0.72

POSITIVE LOGITS

—

0.78

mere

0.78

 equivalents

0.69

 ones

0.64

imental

0.62

 Brilliant

0.60

 Girls

0.59

 Bright

0.58

markets

0.58

 Gand

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.