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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

afort

-0.76

asus

-0.75

folios

-0.74

athan

-0.74

raid

-0.70

otyp

-0.70

agonist

-0.70

agen

-0.67

QB

-0.67

sync

-0.67

POSITIVE LOGITS

 stocking

0.76

 pupils

0.71

Zub

0.67

 commenting

0.65

 foreigners

0.63

 repeating

0.62

 retaining

0.62

 admitting

0.62

 deported

0.61

 discriminated

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.