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

esses

-0.98

ibble

-0.84

oids

-0.83

inker

-0.80

ynthesis

-0.79

iston

-0.77

naissance

-0.75

oop

-0.75

umpy

-0.75

umbers

-0.74

POSITIVE LOGITS

Ved

0.76

 Salvation

0.68

 Georgian

0.67

 vulner

0.67

 Kurd

0.66

 Zoro

0.65

 Kabul

0.64

 truce

0.63

phrine

0.63

 Afghan

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.