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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

elsen

-0.88

roe

-0.72

essen

-0.71

 liqu

-0.71

ternity

-0.69

mington

-0.68

 dipping

-0.68

 amph

-0.68

quist

-0.67

 ropes

-0.67

POSITIVE LOGITS

 Â»

0.67

 Responsibility

0.62

Fay

0.62

Jay

0.61

 Sergei

0.61

rav

0.60

 Â¶

0.60

 Khan

0.59

CHR

0.59

=>

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.