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

pled

-0.94

EStream

-0.84

catentry

-0.83

Ĥ¬

-0.77

cember

-0.76

bucks

-0.71

 practition

-0.71

Ĥª

-0.70

fam

-0.70

eryl

-0.69

POSITIVE LOGITS

iments

0.75

 Stadium

0.75

ers

0.72

imentary

0.71

 Canal

0.65

ances

0.65

 exits

0.65

 GÃ¶

0.65

ings

0.64

 tunnels

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.