INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 Beckham

-0.82

irgin

-0.72

illi

-0.65

 Richards

-0.63

agra

-0.63

anchester

-0.63

daq

-0.62

 Smoking

-0.62

Rey

-0.61

 Hyde

-0.61

POSITIVE LOGITS

 rall

0.69

 Macedonia

0.69

resy

0.68

neck

0.67

inis

0.65

 millenn

0.65

ModLoader

0.64

 Tactics

0.64

pc

0.64

 troop

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.