INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

gr

-0.80

 Brees

-0.76

 Franchise

-0.73

iably

-0.69

 Pierce

-0.69

kins

-0.67

ksh

-0.67

isc

-0.67

phies

-0.66

 Willis

-0.65

POSITIVE LOGITS

 misunder

0.72

alities

0.72

 Kazakh

0.71

ARC

0.70

Reviewer

0.68

SPONSORED

0.66

Rog

0.66

rontal

0.65

 anesthesia

0.65

 goblins

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.