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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

irement

-0.75

itives

-0.72

zbollah

-0.71

 Bonus

-0.65

 resumes

-0.64

hester

-0.64

isites

-0.62

 substitutes

-0.60

 Qual

-0.60

 creditors

-0.60

POSITIVE LOGITS

1.02

inventoryQuantity

0.77

SPONSORED

0.76

éĥ

0.76

cause

0.72

 Archdemon

0.71

alter

0.69

 Cause

0.68

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

0.68

Cause

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.