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

 conviction

-0.63

 Hera

-0.63

-+-+-+-+

-0.61

 framing

-0.60

PID

-0.60

 bond

-0.60

OD

-0.59

 defective

-0.59

cluding

-0.59

 unresolved

-0.59

POSITIVE LOGITS

 laun

0.98

SPONSORED

0.96

iatrics

0.78

Virgin

0.77

DragonMagazine

0.77

 Trop

0.75

aires

0.75

ombat

0.75

âĹ¼

0.75

asher

0.75

Activations Density 0.000%

No Known Activations

This feature has no known activations.