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

 Minor

-0.70

 Reload

-0.66

uld

-0.66

LED

-0.62

 Weasley

-0.62

izoph

-0.60

ufact

-0.60

viation

-0.60

bler

-0.59

 WARN

-0.59

POSITIVE LOGITS

ãĤ¼ãĤ¦ãĤ¹

0.72

 Istanbul

0.71

earable

0.69

PsyNetMessage

0.69

advertisement

0.68

ãĥ¼ãĤ¯

0.66

LinkedIn

0.66

 Bethlehem

0.64

 shaping

0.63

til

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.