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

natureconservancy

-0.93

tnc

-0.68

 padded

-0.67

 juices

-0.66

ippi

-0.66

igham

-0.65

 streng

-0.65

tackle

-0.64

 submar

-0.63

uff

-0.63

POSITIVE LOGITS

JUST

0.77

ieth

0.71

Nou

0.70

idency

0.69

nos

0.69

own

0.69

Nos

0.67

 Micha

0.65

 veter

0.65

NER

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.