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

uyomi

-0.93

nces

-0.78

intendent

-0.63

Neigh

-0.63

æĢ

-0.62

ĳå£«

-0.62

sylv

-0.62

ocene

-0.60

 arous

-0.60

 occurrence

-0.58

POSITIVE LOGITS

 Fiat

0.72

lot

0.72

erity

0.70

milo

0.69

rahim

0.68

ourage

0.67

arin

0.66

 quarters

0.66

 Xiaomi

0.64

rule

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.