INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

fur

-0.73

Kob

-0.70

plet

-0.68

otle

-0.66

scope

-0.65

nee

-0.63

Mexico

-0.63

kaya

-0.61

Registered

-0.60

ception

-0.60

POSITIVE LOGITS

obyl

0.84

 sqor

0.76

 sensit

0.73

 Rahul

0.72

iew

0.71

 unfocusedRange

0.70

 Cumm

0.70

avan

0.68

é¾įåĸļå£«

0.64

 attRot

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.