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

UA

-0.81

oice

-0.79

anny

-0.78

ĸļ

-0.77

avia

-0.74

ender

-0.73

elta

-0.73

armac

-0.72

usky

-0.72

XXX

-0.71

POSITIVE LOGITS

 delusional

0.72

 foregoing

0.71

dil

0.70

 prejud

0.70

 doub

0.69

 understatement

0.67

 hurdles

0.67

 elev

0.64

 impair

0.64

 worlds

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.