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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ahime

-0.77

 VICE

-0.75

 GOODMAN

-0.74

é¾

-0.71

sis

-0.71

rawdownloadcloneembedreportprint

-0.70

 CODE

-0.70

PsyNetMessage

-0.68

CLS

-0.67

Obj

-0.66

POSITIVE LOGITS

 antioxid

0.60

 lieu

0.58

 metabolism

0.58

 reacting

0.57

opausal

0.57

 moistur

0.57

 moist

0.56

arding

0.54

arded

0.54

 bere

0.54

Activations Density 0.000%

No Known Activations

This feature has no known activations.