INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

é¾įå¥ĳå£«

-0.74

edom

-0.73

ovo

-0.71

erity

-0.70

ankind

-0.70

erest

-0.68

oÄŁ

-0.68

 subsequ

-0.65

 actionGroup

-0.65

tsky

-0.64

POSITIVE LOGITS

Zi

0.72

Yin

0.68

zig

0.63

0.60

 Orioles

0.60

 fireworks

0.60

 scrimmage

0.59

aldo

0.59

Bridge

0.58

UI

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.