INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Hungry

-0.69

 favourites

-0.65

Topics

-0.64

DT

-0.62

BCE

-0.62

ocrates

-0.61

 congratulations

-0.60

 respects

-0.60

 curly

-0.60

 diabetic

-0.59

POSITIVE LOGITS

ItemTracker

0.81

unker

0.78

Osw

0.77

umblr

0.71

acters

0.71

heon

0.70

udd

0.69

ouf

0.69

vana

0.68

avia

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.