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

Negative Logits

 Leth

-0.76

Gle

-0.75

 Bound

-0.74

Ĵ

-0.68

 Albion

-0.65

 Esther

-0.64

Nug

-0.64

 Anger

-0.63

verning

-0.62

tal

-0.62

POSITIVE LOGITS

chwitz

0.87

Quote

0.66

psons

0.65

 beverage

0.63

HCR

0.62

Streamer

0.62

Crunch

0.61

chini

0.61

fps

0.61

SourceFile

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.