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

NESS

-0.80

achev

-0.75

merce

-0.72

axter

-0.72

anchester

-0.70

 reinstated

-0.70

 cess

-0.67

anwhile

-0.67

iasco

-0.67

ync

-0.67

POSITIVE LOGITS

ICK

0.71

Ad

0.71

igraph

0.70

Individual

0.69

ocrin

0.66

iology

0.66

HEAD

0.66

sexual

0.66

Table

0.65

Numbers

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.