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

nda

-0.81

 Vendor

-0.76

isSpecialOrderable

-0.75

CVE

-0.73

echo

-0.72

 Invalid

-0.71

 Trials

-0.70

rio

-0.65

dict

-0.65

{"

-0.64

POSITIVE LOGITS

 nipples

0.67

usalem

0.66

 cran

0.66

outhern

0.66

orically

0.64

 Kling

0.63

 circumcised

0.63

eness

0.62

 gland

0.62

 bead

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.