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

Negative Logits

anka

-0.74

 pizz

-0.74

 armour

-0.70

iott

-0.69

 acidic

-0.68

 Predators

-0.68

 spoil

-0.67

 vortex

-0.66

 Ukrainian

-0.66

 cereal

-0.63

POSITIVE LOGITS

ÄŁ

0.78

ching

0.77

ems

0.74

Introduced

0.73

åº

0.70

*=-

0.69

FB

0.68

hao

0.68

 Huang

0.68

ĸļ

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.