INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

recy

-0.71

LECT

-0.71

Ħ¢

-0.70

WAYS

-0.67

 hairs

-0.65

auts

-0.65

REF

-0.65

ItemImage

-0.63

-0.62

POSITIVE LOGITS

accompanied

0.82

rouse

0.77

estyles

0.73

illac

0.72

ciating

0.71

chie

0.68

eer

0.66

occ

0.65

successfully

0.63

ggles

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.