INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Reprodu

-0.77

 Pastebin

-0.72

 reproductive

-0.66

Oro

-0.66

 impe

-0.66

 opio

-0.65

imentary

-0.63

 mosa

-0.62

 environmentalists

-0.62

 reperc

-0.61

POSITIVE LOGITS

rael

0.85

é¾į

0.83

obook

0.81

Reviewer

0.75

wow

0.73

å¥³

0.71

Ðº

0.70

adjusted

0.69

ahime

0.69

PET

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.