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

 sabotage

-0.77

 spont

-0.71

upon

-0.69

 compromises

-0.68

 sanity

-0.67

omach

-0.65

icker

-0.64

 derail

-0.64

-0.63

 surprises

-0.62

POSITIVE LOGITS

ources

0.72

 Throne

0.69

 Inher

0.66

Services

0.65

 Coffin

0.64

andowski

0.64

 Zucker

0.64

phthal

0.63

 Counsel

0.63

 Gifts

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.