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

 Giuliani

-0.80

opes

-0.75

iaz

-0.75

edu

-0.71

oks

-0.71

umerable

-0.70

ichita

-0.69

iture

-0.68

izer

-0.67

ulin

-0.66

POSITIVE LOGITS

 Shutdown

0.72

 Coffin

0.69

accompan

0.65

OY

0.64

 sustaining

0.63

Â¯Â¯

0.63

 dere

0.62

lat

0.62

 blockade

0.62

Crew

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.