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

quickShipAvailable

-0.86

matter

-0.82

©¶æ

-0.79

ledged

-0.77

 withd

-0.75

etheless

-0.75

ugu

-0.75

wcsstore

-0.74

Ļ½

-0.74

waves

-0.73

POSITIVE LOGITS

ACP

0.71

ific

0.66

0.64

 GOODMAN

0.63

 Parm

0.63

 gallons

0.62

 gallon

0.62

AC

0.62

STL

0.61

 stud

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.