INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Insp

-0.81

vot

-0.73

ItemTracker

-0.72

 Sets

-0.68

Pil

-0.65

jew

-0.64

 catches

-0.63

 Bangl

-0.62

 Contest

-0.62

 Mighty

-0.62

POSITIVE LOGITS

nee

0.76

ppard

0.75

uter

0.74

rup

0.70

ibia

0.70

enburg

0.69

rib

0.69

walk

0.68

rive

0.68

fee

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.