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

hall

-0.84

WARE

-0.70

eer

-0.69

dar

-0.69

Kings

-0.68

Eng

-0.67

anguage

-0.66

nas

-0.66

oyer

-0.65

Ed

-0.64

POSITIVE LOGITS

 tatt

0.86

 bouncing

0.67

Zup

0.64

 snatch

0.63

TTC

0.63

 blasts

0.61

 retard

0.60

 torpedo

0.58

æŃ

0.58

 targ

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.