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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

.","

-0.69

ilege

-0.68

":{"

-0.67

 reservation

-0.67

asis

-0.66

icho

-0.66

 Compatibility

-0.65

osis

-0.65

rogens

-0.64

adoes

-0.64

POSITIVE LOGITS

SHIP

0.75

EY

0.67

âĸ¬âĸ¬

0.65

LEY

0.64

EVA

0.63

Python

0.63

AMA

0.61

PIN

0.61

Ana

0.61

WOOD

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.