INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 mund

-0.64

endif

-0.62

fac

-0.61

rane

-0.61

 Admir

-0.59

ukong

-0.59

 Virtue

-0.59

olt

-0.59

dit

-0.59

riv

-0.59

POSITIVE LOGITS

 misc

0.63

achu

0.62

 chops

0.57

fights

0.56

alach

0.56

 transition

0.55

enum

0.55

 successfully

0.54

 matches

0.54

wal

0.54

Activations Density 0.000%

No Known Activations

This feature has no known activations.