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

schild

-0.68

rees

-0.67

sonian

-0.67

quished

-0.65

 mourn

-0.65

ury

-0.64

irl

-0.63

elight

-0.63

erred

-0.63

istrates

-0.63

POSITIVE LOGITS

catentry

0.70

Kah

0.61

quickShipAvailable

0.60

assian

0.60

 Uganda

0.60

>"

0.59

 breaching

0.58

CLASSIFIED

0.58

 discrep

0.58

Card

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.