INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

raised

-0.87

locking

-0.83

mental

-0.81

translation

-0.79

assisted

-0.74

pressed

-0.74

shot

-0.73

fetched

-0.73

blown

-0.72

locks

-0.72

POSITIVE LOGITS

 Journals

0.83

 Vide

0.75

senal

0.71

VID

0.71

 OFFIC

0.66

 aloud

0.66

 Spect

0.65

 english

0.64

 Volcano

0.64

 Beaut

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.