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

testing

-0.72

chrom

-0.67

 ILCS

-0.65

angular

-0.64

cro

-0.62

spect

-0.60

bows

-0.58

xes

-0.58

COL

-0.58

govtrack

-0.58

POSITIVE LOGITS

milo

0.79

 Harley

0.71

ensable

0.70

rete

0.67

ensity

0.66

roups

0.66

rief

0.65

ften

0.64

oustic

0.64

Cooldown

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.