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

thood

-0.68

ational

-0.67

iors

-0.65

 Hearts

-0.63

arat

-0.61

Gym

-0.61

 Recreation

-0.60

 inherited

-0.60

 Thor

-0.59

 Amph

-0.59

POSITIVE LOGITS

rites

0.69

 brim

0.65

Ħ¢

0.65

 dich

0.64

TPS

0.63

mouth

0.62

utherland

0.59

aturday

0.59

DAQ

0.58

Freedom

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.