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

inders

-0.80

tackle

-0.75

 retract

-0.69

vals

-0.68

 SPORTS

-0.68

DL

-0.67

 BALL

-0.66

igham

-0.65

 Flags

-0.64

 vetoed

-0.63

POSITIVE LOGITS

 Plague

0.77

 Tart

0.74

acid

0.74

 enthusi

0.73

Sym

0.70

ety

0.68

quartered

0.68

Ec

0.68

 itch

0.66

complex

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.