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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

:{

-0.84

swick

-0.79

ullah

-0.73

 forwards

-0.72

ister

-0.69

¢

-0.68

chens

-0.66

orge

-0.63

 Hicks

-0.62

contact

-0.62

POSITIVE LOGITS

naissance

0.79

 resil

0.76

ipop

0.73

LCS

0.68

 infringing

0.68

ISC

0.67

activation

0.65

 Wanted

0.63

isoft

0.63

ado

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.