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

ngth

-0.58

elaide

-0.58

 newsletter

-0.58

mberg

-0.57

door

-0.57

ream

-0.57

 Feld

-0.56

catentry

-0.55

eland

-0.55

POSITIVE LOGITS

inav

0.77

ammad

0.68

anson

0.68

 Warfare

0.65

Tags

0.64

pard

0.63

Heb

0.62

apy

0.62

¬¼

0.61

esis

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.