INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Piano

-0.71

 Kobe

-0.66

 Clover

-0.64

 Precision

-0.63

 arbitration

-0.63

GS

-0.62

 sealed

-0.62

SF

-0.61

SF

-0.61

CLE

-0.61

POSITIVE LOGITS

idas

0.69

bedroom

0.64

 incumb

0.64

discrimination

0.64

htaking

0.60

rimination

0.59

ocracy

0.59

geant

0.59

 Dres

0.59

 distingu

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.