INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Ames

-0.79

hesda

-0.70

------------------------------------------------

-0.70

 indemn

-0.68

Bey

-0.67

ukong

-0.64

alse

-0.64

 Aval

-0.62

 Arche

-0.60

 Lowell

-0.60

POSITIVE LOGITS

oir

0.66

rack

0.64

SG

0.62

 puberty

0.61

orr

0.61

oga

0.61

wagen

0.61

çİĭ

0.60

 gorilla

0.59

GBT

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.