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

nir

-0.84

ANK

-0.82

esta

-0.79

heddar

-0.77

flies

-0.73

bush

-0.72

Greek

-0.71

 Yamato

-0.69

rain

-0.69

¿

-0.68

POSITIVE LOGITS

eatures

0.86

 boycott

0.84

eport

0.71

 refrain

0.68

 discourse

0.65

 recreation

0.65

 abstinence

0.64

ruciating

0.63

 clinch

0.63

 impossibility

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.