INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

leigh

-0.78

\\\\\\\\

-0.78

tions

-0.74

eteenth

-0.73

ãĥ´

-0.71

endez

-0.70

querque

-0.69

CHAT

-0.69

los

-0.68

 condem

-0.68

POSITIVE LOGITS

 Marketable

0.77

milo

0.68

ptin

0.67

hidden

0.63

 leptin

0.63

 lean

0.61

utive

0.60

 surg

0.59

netflix

0.59

acus

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.