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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

utral

-0.74

fters

-0.68

nce

-0.68

 Assist

-0.61

 phased

-0.61

ner

-0.61

cess

-0.59

eele

-0.57

 immersion

-0.57

 upgr

-0.57

POSITIVE LOGITS

odynam

0.74

kefeller

0.72

cul

0.65

anmar

0.64

OHN

0.63

laws

0.63

ITH

0.63

 cannabin

0.63

abouts

0.63

 paternity

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.