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

oun

-0.78

otaur

-0.74

©¶æ

-0.73

robat

-0.72

parts

-0.68

 Aires

-0.67

urances

-0.67

 Extras

-0.66

 Patriarch

-0.65

©¶æ¥µ

-0.65

POSITIVE LOGITS

CVE

0.66

lear

0.64

dh

0.64

FC

0.64

isSpecialOrderable

0.63

Dys

0.57

le

0.56

'';

0.56

afer

0.55

awar

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.