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

ħĭ

-0.84

uctions

-0.72

BUG

-0.70

»Ĵ

-0.70

 Liang

-0.69

©¶æ¥µ

-0.69

©¶æ

-0.67

EStream

-0.67

 caval

-0.66

ª

-0.63

POSITIVE LOGITS

hered

0.76

Suc

0.72

Registered

0.70

recy

0.70

blogspot

0.70

successfully

0.68

aligned

0.68

atform

0.67

dated

0.66

wired

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.