INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

acci

-0.75

ulty

-0.74

 Primary

-0.73

mson

-0.68

etsk

-0.61

 Alps

-0.61

 Briggs

-0.60

regor

-0.59

artifacts

-0.59

born

-0.58

POSITIVE LOGITS

xit

0.87

ÙĴ

0.85

 è£ıè

0.76

uces

0.76

ãģĨ

0.75

Hat

0.74

ãĥĥãĥī

0.74

Fat

0.72

Ô

0.70

ãĥ£

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.