INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

azi

-0.73

ument

-0.72

gate

-0.71

raine

-0.66

 Stand

-0.63

ija

-0.62

 captcha

-0.60

 aisle

-0.59

 psychosis

-0.59

cheon

-0.58

POSITIVE LOGITS

pmwiki

0.79

bryce

0.70

dit

0.68

amara

0.67

 LIMITED

0.66

dump

0.63

erenn

0.63

shall

0.63

theless

0.62

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.