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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

conservancy

-0.72

pract

-0.68

ĻĤ

-0.67

ometers

-0.66

 trump

-0.66

 Muslims

-0.65

Muslims

-0.63

asks

-0.63

 Ivanka

-0.62

ocrats

-0.62

POSITIVE LOGITS

tti

0.82

WD

0.69

iton

0.68

dos

0.67

ãĥ´

0.65

 promot

0.65

 WRITE

0.62

DX

0.61

 Rodrigo

0.61

 Leone

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.