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

RD

-0.74

DN

-0.72

ND

-0.70

wx

-0.67

gd

-0.67

Ã¦

-0.66

WT

-0.66

Dol

-0.66

 Nationals

-0.66

ansk

-0.66

POSITIVE LOGITS

 fusion

0.75

ŃĶ

0.71

chnology

0.70

Ħ¢

0.67

 scra

0.66

»Ĵ

0.63

tut

0.63

ģĸ

0.61

 tink

0.60

 tack

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.