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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Jihad

-0.73

 Meditation

-0.73

UAL

-0.70

¶æ

-0.69

 Terror

-0.67

 Aura

-0.66

 Scholars

-0.65

POSE

-0.65

vid

-0.64

è

-0.64

POSITIVE LOGITS

hran

0.78

ufact

0.78

 soda

0.77

 conditioned

0.72

 compressor

0.71

acco

0.69

 piping

0.69

riet

0.69

sylv

0.69

cil

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.