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

ahime

-0.83

Downloadha

-0.73

 Celt

-0.73

ocity

-0.73

 Cosponsors

-0.70

coe

-0.68

à©

-0.66

UGH

-0.65

bol

-0.65

ordes

-0.63

POSITIVE LOGITS

 Skydragon

0.74

 organising

0.64

plates

0.61

 charity

0.61

 Pist

0.61

 earners

0.60

illon

0.60

 heaven

0.60

 inheritance

0.59

pher

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.