INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

olding

-0.74

olds

-0.71

rafted

-0.68

stocks

-0.66

©¶æ

-0.63

 Ital

-0.62

 Blazing

-0.60

ailability

-0.60

ounds

-0.60

 appearing

-0.59

POSITIVE LOGITS

gow

0.76

achu

0.65

oway

0.65

ussian

0.61

kay

0.61

gerald

0.60

alog

0.60

hari

0.60

uss

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.