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

trl

-0.79

 Shack

-0.71

oded

-0.70

solete

-0.70

odore

-0.69

numbered

-0.69

collar

-0.68

 Lantern

-0.67

 tradem

-0.66

ramid

-0.66

POSITIVE LOGITS

––

0.66

steam

0.65

 Bian

0.64

 inval

0.64

eers

0.63

eu

0.62

fur

0.62

ealous

0.60

ele

0.60

IENCE

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.