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

tch

-0.77

REAM

-0.75

hops

-0.75

lass

-0.75

hots

-0.75

onde

-0.74

perse

-0.73

hower

-0.73

ican

-0.72

heric

-0.72

POSITIVE LOGITS

 Canaver

0.70

 practitioners

0.67

 unions

0.66

 Cure

0.63

endor

0.62

 stigma

0.61

gp

0.61

 advent

0.59

 union

0.59

 Cups

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.