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

 Reloaded

-0.69

 Tactics

-0.64

 Telesc

-0.63

 travels

-0.60

 emanc

-0.59

 Redemption

-0.58

maxwell

-0.57

 Shattered

-0.57

 Pressure

-0.56

 immersion

-0.55

POSITIVE LOGITS

Ħ

0.76

¬¼

0.74

umeric

0.69

 watchdog

0.69

Ģ

0.68

ika

0.66

FEC

0.66

anie

0.66

 Sabha

0.65

ĺħ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.