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

 Invention

-0.71

 Franch

-0.65

 Horses

-0.64

Ard

-0.63

 Scenes

-0.63

 Shine

-0.61

 Uber

-0.61

 Speed

-0.60

 namely

-0.59

 Canaan

-0.58

POSITIVE LOGITS

illes

0.78

agascar

0.76

Wan

0.74

ushima

0.71

icans

0.70

isf

0.69

oma

0.69

ogl

0.69

 indebted

0.68

odon

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.