INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ike

-0.77

fman

-0.74

ikes

-0.73

ona

-0.72

uns

-0.71

mort

-0.67

elo

-0.66

asta

-0.66

III

-0.66

bara

-0.66

POSITIVE LOGITS

 warr

0.80

 Compos

0.72

Sov

0.68

 Pradesh

0.67

ãĥ´

0.66

 defic

0.66

 Rhino

0.65

 à¨

0.63

 à¤

0.63

lict

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.