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

glers

-0.74

 comed

-0.67

 Ulster

-0.64

 Vine

-0.64

aceous

-0.62

 Polic

-0.62

 Soldiers

-0.61

Pv

-0.61

 Gorsuch

-0.61

 Sorce

-0.60

POSITIVE LOGITS

athy

0.87

unts

0.74

anton

0.71

independence

0.71

kus

0.71

ngth

0.70

ector

0.69

xon

0.69

fman

0.68

eyes

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.