INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

uther

-0.74

 ceilings

-0.72

ali

-0.70

 horsepower

-0.68

osity

-0.65

 bulb

-0.64

Redditor

-0.64

dump

-0.64

reddits

-0.64

 dues

-0.63

POSITIVE LOGITS

 Defenders

0.80

axter

0.77

FORE

0.74

 Dominion

0.73

 Gates

0.67

Cas

0.66

 Vanguard

0.66

 Tribune

0.65

BP

0.64

 Marriott

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.