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

 closet

-0.81

gaard

-0.69

 boxer

-0.67

 detox

-0.67

 chemotherapy

-0.66

 surfing

-0.65

 locker

-0.64

 pregnancy

-0.63

 cycling

-0.63

 furnace

-0.62

POSITIVE LOGITS

hens

0.83

rots

0.74

omen

0.72

iris

0.69

itual

0.68

ASC

0.67

Fine

0.66

ensional

0.65

yright

0.65

 Fancy

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.