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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gri

-0.66

ashington

-0.65

 guid

-0.65

street

-0.64

 Templ

-0.62

Street

-0.61

ERSON

-0.61

.""

-0.61

agall

-0.60

 psychiat

-0.59

POSITIVE LOGITS

zai

0.74

Uk

0.74

nings

0.69

meal

0.65

relevant

0.62

apult

0.61

dist

0.61

acly

0.61

nown

0.61

gif

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.