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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Baal

-0.88

repeat

-0.75

reon

-0.74

Cao

-0.72

pur

-0.67

Demon

-0.67

die

-0.66

ser

-0.65

Joined

-0.65

Thu

-0.65

POSITIVE LOGITS

jri

0.78

 Journals

0.68

GPA

0.65

leeve

0.64

 Cabinet

0.64

lement

0.63

poons

0.62

icultural

0.62

 gallons

0.62

FFER

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.