INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

arthed

-0.69

tumblr

-0.69

Applic

-0.68

sbm

-0.67

aturdays

-0.65

swer

-0.64

APD

-0.62

EXP

-0.62

Past

-0.62

zona

-0.62

POSITIVE LOGITS

—

0.76

ãĤ¨ãĥ«

0.75

rots

0.73

—-

0.68

—"

0.65

"—

0.64

,—

0.64

atar

0.63

dinand

0.62

 declass

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.