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

izabeth

-0.75

âĨĳ

-0.72

 Huck

-0.71

CLASSIFIED

-0.68

 Aden

-0.68

 Tycoon

-0.68

emo

-0.67

href

-0.66

 Calais

-0.65

TOP

-0.65

POSITIVE LOGITS

urus

0.70

animous

0.70

inguished

0.69

ixture

0.66

ocious

0.65

aber

0.64

 inevitable

0.63

culus

0.63

common

0.62

osuke

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.