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

uckland

-0.97

aeda

-0.84

£ı

-0.76

thora

-0.70

ayne

-0.70

zsche

-0.69

ench

-0.67

Jew

-0.66

aqu

-0.66

oteric

-0.65

POSITIVE LOGITS

 Papers

0.69

 preparation

0.64

rant

0.63

erent

0.63

0.62

 firsthand

0.61

 memoir

0.60

 Peyton

0.59

 Peggy

0.58

ctic

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.