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

 Pigs

-0.68

(*

-0.62

 Salt

-0.61

Target

-0.58

iment

-0.58

 worm

-0.56

Salt

-0.56

 injection

-0.56

CEPT

-0.56

Training

-0.56

POSITIVE LOGITS

oulos

0.84

ourgeois

0.83

ideshow

0.83

merce

0.81

 partName

0.79

ORGE

0.78

ibli

0.77

oud

0.76

vir

0.75

aston

0.75

Activations Density 0.000%

No Known Activations

This feature has no known activations.