INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

UAL

-0.76

 overlap

-0.71

nikov

-0.70

 beard

-0.66

Rating

-0.65

iosyncr

-0.65

moon

-0.64

leeve

-0.63

 defic

-0.63

POSITIVE LOGITS

warts

0.74

ercise

0.72

renheit

0.70

icol

0.68

 academ

0.68

 Hogwarts

0.67

apon

0.66

urses

0.65

asus

0.64

 Galile

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.