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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

leneck

-0.69

 infiltrated

-0.66

hid

-0.65

NAS

-0.63

ãĥĨ

-0.63

 Haas

-0.63

iosyn

-0.62

alore

-0.62

vantage

-0.62

alysis

-0.60

POSITIVE LOGITS

ials

0.77

oct

0.70

rose

0.69

 numb

0.68

amaz

0.66

XXX

0.65

rase

0.65

amaru

0.65

aneous

0.64

iable

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.