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

tein

-0.83

xtap

-0.78

pton

-0.76

schild

-0.76

ensibly

-0.75

apo

-0.70

ospace

-0.69

ocally

-0.69

Wiki

-0.68

wolf

-0.68

POSITIVE LOGITS

 Advisory

0.78

 Aval

0.73

 Charge

0.72

 Technician

0.68

 Reason

0.66

iT

0.60

 Guards

0.59

 Imper

0.59

 Intervention

0.59

 Depths

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.