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

ĨĴ

-0.79

¿½

-0.73

onica

-0.70

aughtered

-0.69

uba

-0.65

 Sparks

-0.61

 Rhythm

-0.60

 Burton

-0.60

 Dickinson

-0.60

 Knock

-0.59

POSITIVE LOGITS

=]

0.82

Ô

0.69

":"/

0.65

umbai

0.64

 remem

0.64

0.63

Ã¦

0.62

emet

0.62

devices

0.61

à©

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.