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

 digs

-0.70

 Ammo

-0.66

!/

-0.64

 Immunity

-0.63

 alas

-0.60

 accol

-0.59

 vulner

-0.59

 confines

-0.59

 toget

-0.59

 hires

-0.58

POSITIVE LOGITS

redd

0.73

entary

0.72

Ļ½

0.71

larg

0.71

notation

0.71

itia

0.69

alky

0.69

 Noir

0.68

cham

0.68

gres

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.