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

shapeshifter

-0.76

orum

-0.75

aughtered

-0.74

ildo

-0.66

gas

-0.65

ionic

-0.64

endon

-0.64

 Pegasus

-0.64

>>\

-0.63

usky

-0.61

POSITIVE LOGITS

renheit

0.77

soever

0.76

xtap

0.73

witz

0.72

ITED

0.69

ãĤ¹ãĥĪ

0.68

 Townsend

0.67

avorite

0.67

 Tenn

0.66

Town

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.