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

 needles

-0.73

oward

-0.63

 applic

-0.60

 readings

-0.59

 provocation

-0.59

 shoots

-0.59

 playbook

-0.58

 orally

-0.58

 capsule

-0.58

 congratulations

-0.57

POSITIVE LOGITS

athered

0.80

Mesh

0.78

spin

0.75

imum

0.73

 guiActiveUnfocused

0.72

sv

0.69

tions

0.68

erred

0.67

FFFF

0.66

ilib

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.