INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

otine

-0.76

 exerc

-0.74

alysed

-0.73

 smelling

-0.71

 toile

-0.69

emark

-0.68

 paras

-0.65

isphere

-0.63

 purse

-0.63

ijn

-0.62

POSITIVE LOGITS

doms

0.76

WAR

0.73

uala

0.68

Leaks

0.68

ysical

0.68

gs

0.62

ankind

0.62

jong

0.62

missions

0.61

aven

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.