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

qus

-0.74

 tremend

-0.71

ĸļ

-0.67

 Zucker

-0.67

onga

-0.66

£ı

-0.66

 domestically

-0.66

 embry

-0.63

 wont

-0.63

MFT

-0.62

POSITIVE LOGITS

fitting

0.82

andals

0.68

erness

0.67

ortment

0.65

 Volunte

0.60

etry

0.59

 cracks

0.59

intage

0.58

pees

0.58

yards

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.