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

lust

-0.74

jad

-0.73

calling

-0.73

UTERS

-0.73

 Gleaming

-0.73

bats

-0.71

Sad

-0.71

zeb

-0.69

AUT

-0.69

Austral

-0.68

POSITIVE LOGITS

 accomp

0.75

hetical

0.71

orman

0.66

 repr

0.65

ausible

0.63

 stake

0.62

 Lever

0.62

ciplinary

0.62

usercontent

0.62

iosyncr

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.