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

umen

-0.68

 brim

-0.66

 needles

-0.66

Oz

-0.66

!]

-0.64

erella

-0.63

 Monstrous

-0.62

 dwarves

-0.62

pee

-0.62

utils

-0.61

POSITIVE LOGITS

externalActionCode

0.92

anca

0.83

ITIES

0.65

avascript

0.65

Mobile

0.65

together

0.64

angan

0.63

iciary

0.62

acqu

0.61

where

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.