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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.83

 Franch

-0.76

 Kamp

-0.68

initions

-0.68

pun

-0.67

 tables

-0.65

quickShipAvailable

-0.65

avor

-0.65

 Grimoire

-0.64

table

-0.64

POSITIVE LOGITS

arya

0.72

 digit

0.72

 nucleus

0.69

erg

0.67

cs

0.65

hart

0.65

 Â·

0.64

 impacted

0.64

 quake

0.63

 recovery

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.