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

 subclass

-0.72

LET

-0.68

Statement

-0.66

Tub

-0.66

INESS

-0.65

 Trad

-0.63

 Remem

-0.62

BIT

-0.61

 Loot

-0.60

OLOGY

-0.60

POSITIVE LOGITS

ester

1.50

arten

0.85

esters

0.79

itol

0.75

istant

0.70

rha

0.70

_>

0.69

aults

0.68

ior

0.68

ento

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.