INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

hift

-0.79

atform

-0.72

racted

-0.71

ilater

-0.70

marked

-0.68

shift

-0.68

ometimes

-0.67

xton

-0.65

STRUCT

-0.64

division

-0.64

POSITIVE LOGITS

0.80

æ©Ł

0.78

ashington

0.76

çīĪ

0.71

 condem

0.68

Klu

0.67

ãĤ¨ãĥ«

0.65

Nad

0.65

dos

0.65

 shoots

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.