INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 tradem

-0.81

 encount

-0.74

senal

-0.72

ccording

-0.71

edIn

-0.71

 hemor

-0.69

PDATE

-0.67

ELD

-0.65

estead

-0.65

erest

-0.64

POSITIVE LOGITS

 Planes

0.74

anon

0.70

itbart

0.69

 Weasley

0.69

cn

0.69

rage

0.68

ban

0.68

myra

0.67

mare

0.67

baum

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.