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

 faults

-0.73

Dou

-0.68

 Yiannopoulos

-0.67

uries

-0.66

":[

-0.64

 slams

-0.60

{:

-0.60

 caveats

-0.59

ãĤº

-0.58

MX

-0.57

POSITIVE LOGITS

avin

0.69

ModLoader

0.68

SHIP

0.66

awei

0.65

0.64

luster

0.63

mitt

0.61

ogan

0.61

Þ

0.60

 withd

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.