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

clusive

-0.77

 integ

-0.68

 explor

-0.66

 seed

-0.65

 isot

-0.62

akin

-0.61

 astronomical

-0.61

seed

-0.60

enture

-0.60

urally

-0.60

POSITIVE LOGITS

aign

0.97

 Flavoring

0.79

ãĥ´

0.75

yu

0.68

irez

0.68

ghazi

0.66

incible

0.64

bilt

0.64

ught

0.62

rez

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.