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

iliary

-0.79

lag

-0.79

rences

-0.75

egg

-0.73

plings

-0.72

agame

-0.71

ilings

-0.71

imum

-0.70

Ã

-0.67

oldown

-0.67

POSITIVE LOGITS

 Anyway

0.70

 Scal

0.68

gun

0.68

 Oath

0.67

 privileged

0.65

pa

0.63

 moss

0.63

 sacks

0.61

 unmarked

0.61

ticket

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.