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

icative

-0.71

uble

-0.71

}\

-0.68

)]

-0.68

eling

-0.67

hap

-0.67

(/

-0.65

nel

-0.64

urat

-0.63

jri

-0.63

POSITIVE LOGITS

antis

0.72

 Extras

0.69

esson

0.65

 Cosponsors

0.61

foundland

0.60

notations

0.59

 Hera

0.59

 QUEST

0.58

 indic

0.58

aeper

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.