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

otti

-0.80

assis

-0.69

 flattering

-0.68

 nods

-0.67

ISI

-0.67

idency

-0.66

 sensit

-0.66

ijn

-0.65

 recognition

-0.64

pins

-0.64

POSITIVE LOGITS

Gib

0.74

 Inher

0.70

ä

0.68

hander

0.68

 Swordsman

0.65

eden

0.65

izens

0.64

Ich

0.63

duration

0.63

 hosp

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.