INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

lime

-0.66

ichick

-0.65

lead

-0.62

 è£ıè

-0.60

OC

-0.60

imeter

-0.60

zer

-0.60

aez

-0.60

ubes

-0.60

ME

-0.60

POSITIVE LOGITS

 Cheong

0.86

 Peng

0.84

ishly

0.72

Hai

0.66

avorite

0.66

ccess

0.65

etheless

0.64

 shenan

0.64

osponsors

0.62

Accessory

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.