INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

failed

-0.66

DEM

-0.66

 nominate

-0.63

RY

-0.63

dev

-0.62

Kon

-0.61

Rah

-0.60

Hud

-0.60

 Krypt

-0.60

aj

-0.59

POSITIVE LOGITS

acial

0.79

isure

0.74

 passports

0.72

 ivory

0.71

entimes

0.69

ople

0.68

thren

0.67

isine

0.67

elman

0.66

ertodd

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.