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

 happiest

-0.74

idity

-0.73

erity

-0.67

rity

-0.67

aughs

-0.66

rities

-0.64

 Kinnikuman

-0.63

 Monroe

-0.63

 metic

-0.63

 strongest

-0.62

POSITIVE LOGITS

Blocks

0.85

dj

0.73

we

0.72

Chain

0.72

invoke

0.68

trip

0.68

 Trip

0.67

rip

0.66

Lua

0.63

dan

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.