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

hiba

-0.70

IDES

-0.64

 sequ

-0.63

merce

-0.61

athing

-0.60

iaries

-0.59

raits

-0.59

 Dexter

-0.58

ilts

-0.58

livion

-0.57

POSITIVE LOGITS

achev

0.71

 negotiation

0.64

ams

0.64

aff

0.63

dem

0.63

Ð´

0.62

ilateral

0.62

 Means

0.61

dom

0.60

 persuasion

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.