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

 Scale

-0.69

 Journals

-0.66

 Underground

-0.65

xes

-0.64

 Vehicles

-0.63

 Solid

-0.63

ths

-0.63

amsung

-0.62

 Massive

-0.61

 Spec

-0.61

POSITIVE LOGITS

rika

0.77

oris

0.76

onymous

0.75

oulos

0.74

uki

0.74

nel

0.73

idon

0.72

rank

0.69

bek

0.68

emi

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.