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

 forwards

-0.68

cest

-0.66

icity

-0.66

kinson

-0.63

 guys

-0.62

 descriptions

-0.61

neau

-0.60

ãĥ¥

-0.60

cÃ©

-0.59

 commissions

-0.58

POSITIVE LOGITS

pter

0.90

 indo

0.71

argo

0.69

taboola

0.67

'[

0.66

angible

0.65

rained

0.63

 Sanct

0.62

Sport

0.61

habi

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.