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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 dealers

-0.75

illac

-0.64

 fleets

-0.63

 conditioning

-0.63

iott

-0.62

atos

-0.61

 assists

-0.61

 bombers

-0.61

ansky

-0.60

 tanks

-0.60

POSITIVE LOGITS

 THEN

0.72

ailable

0.70

 DOWN

0.69

TPS

0.68

Ga

0.67

iae

0.65

Gle

0.65

izont

0.65

 DRAG

0.64

hew

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.