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

DAQ

-0.86

nesday

-0.82

thood

-0.81

nces

-0.78

norm

-0.75

etheless

-0.75

orthern

-0.73

antry

-0.73

lease

-0.69

 resil

-0.68

POSITIVE LOGITS

ost

0.70

 GOODMAN

0.66

isman

0.64

 Jugg

0.63

 eyebrow

0.61

seed

0.58

 dismissing

0.58

etry

0.55

 highlights

0.55

 dwar

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.