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

 Carbuncle

-0.91

theless

-0.74

 quake

-0.66

Gou

-0.65

 enthus

-0.64

atform

-0.62

IRC

-0.62

å£«

-0.62

ointed

-0.61

Â¯Â¯

-0.61

POSITIVE LOGITS

 rights

0.81

etr

0.77

nesday

0.77

haul

0.72

 Observatory

0.72

yrights

0.69

ymm

0.69

rolet

0.67

Downloadha

0.62

rey

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.