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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

psc

-0.84

prus

-0.76

ippi

-0.72

xtap

-0.68

bara

-0.67

igham

-0.64

nery

-0.63

bench

-0.63

istries

-0.63

ugen

-0.62

POSITIVE LOGITS

é¾įåĸļå£«

0.74

romising

0.74

 Beasts

0.73

è£ħ

0.67

REDACTED

0.66

ilant

0.63

uca

0.62

 Lear

0.61

 iCloud

0.61

immer

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.