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

 oxid

-0.68

 fasc

-0.68

bourg

-0.67

ymes

-0.67

 corners

-0.66

 intimately

-0.65

 origin

-0.64

 interrupts

-0.63

 phases

-0.63

 circumference

-0.62

POSITIVE LOGITS

DoS

0.84

acia

0.74

papers

0.74

CCC

0.71

abee

0.70

rieve

0.69

CE

0.68

cmp

0.68

AGES

0.67

agher

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.