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

tein

-0.78

Sent

-0.72

Ire

-0.71

Write

-0.68

ãĥ³ãĤ¸

-0.68

ãĥ¼ãĥĨ

-0.66

ESE

-0.65

ãĥĩ

-0.62

IPS

-0.61

Words

-0.60

POSITIVE LOGITS

plex

0.91

Cav

0.74

 Debor

0.71

raught

0.71

akespe

0.69

thening

0.67

erc

0.63

alled

0.62

ellery

0.62

htaking

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.