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

 =================================

-0.67

ancial

-0.65

err

-0.65

GCC

-0.62

orst

-0.62

 endif

-0.62

 secondly

-0.61

FTP

-0.61

olate

-0.61

 --------

-0.60

POSITIVE LOGITS

 accuracy

1.00

ury

0.79

uries

0.78

ãĤ¨ãĥ«

0.77

urers

0.73

Berry

0.72

urance

0.70

ãĤ¦ãĤ¹

0.70

é¾į

0.68

rower

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.