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

 Chocobo

-0.72

>(

-0.69

 plagiar

-0.69

.)

-0.68

 technically

-0.65

idium

-0.64

tein

-0.63

 adam

-0.62

rahim

-0.62

 sexually

-0.61

POSITIVE LOGITS

Rex

0.73

watch

0.69

Shape

0.68

 Comic

0.67

Tact

0.66

Vo

0.63

wra

0.62

osp

0.62

Split

0.62

Dust

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.