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

 Jamaica

-0.72

 Conan

-0.67

 Gibraltar

-0.66

 Hitchcock

-0.65

 Desmond

-0.62

 dere

-0.62

 Cumber

-0.60

 Tarant

-0.59

 Bermuda

-0.59

Dar

-0.59

POSITIVE LOGITS

pole

0.67

hiba

0.66

osaurus

0.66

rehend

0.64

eus

0.64

 pione

0.64

nown

0.64

erva

0.64

alla

0.63

roach

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.