INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

heit

-0.79

 Hazard

-0.67

ãĤ½

-0.66

 bleach

-0.65

isner

-0.65

nos

-0.65

isec

-0.64

Fore

-0.64

Pont

-0.62

 ecology

-0.62

POSITIVE LOGITS

}:

0.66

ãĤ¤ãĥĪ

0.63

bos

0.63

atl

0.62

ANA

0.60

iliate

0.58

Express

0.58

 ends

0.57

otine

0.57

lude

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.