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

morrow

-0.74

mination

-0.73

clair

-0.72

mine

-0.71

hower

-0.68

hani

-0.66

OA

-0.66

heid

-0.66

AIDS

-0.66

bage

-0.65

POSITIVE LOGITS

)</

0.70

 Carbuncle

0.65

 stoked

0.63

oulos

0.61

omatic

0.61

ovy

0.60

ibles

0.60

iatric

0.59

uffed

0.59

 recl

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.