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

keep

-0.67

 Goods

-0.65

iments

-0.59

 Accessories

-0.58

sharing

-0.57

places

-0.57

deep

-0.56

 Feedback

-0.55

 formats

-0.55

ware

-0.55

POSITIVE LOGITS

realDonaldTrump

0.82

ãĥĥãĥī

0.77

ensen

0.72

REDACTED

0.72

Ń·

0.71

INTON

0.70

enson

0.70

å¥³

0.68

nesota

0.67

PAC

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.