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

ggles

-0.73

#$

-0.68

ctic

-0.68

visors

-0.67

 displ

-0.65

 Hopkins

-0.62

intendent

-0.60

acity

-0.60

ãĥ¼ãĥĨ

-0.59

AMES

-0.59

POSITIVE LOGITS

noticed

0.78

sequ

0.76

interstitial

0.75

ertodd

0.72

hess

0.71

"]

0.69

govtrack

0.69

arella

0.64

ore

0.62

bitcoin

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.