INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ractor

-0.82

ickr

-0.78

icket

-0.75

ackle

-0.75

versive

-0.73

renheit

-0.72

 lodge

-0.70

pez

-0.70

 watering

-0.68

rat

-0.67

POSITIVE LOGITS

 Franch

0.73

Ess

0.69

Definition

0.66

 Defin

0.65

ãĤ¼

0.64

nect

0.64

 Stephenson

0.63

RG

0.62

Level

0.61

 Mamm

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.