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

ventus

-0.96

ipeg

-0.86

psey

-0.81

Accessory

-0.78

liga

-0.77

etsk

-0.77

ixtape

-0.73

bery

-0.72

idas

-0.72

edin

-0.71

POSITIVE LOGITS

ENSE

0.63

gerald

0.62

ENC

0.58

ISTORY

0.58

 marrow

0.56

 lifetime

0.56

esan

0.56

 LIFE

0.55

oy

0.55

DOC

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.