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

 manif

-0.78

ynes

-0.71

é¾įåĸļå£«

-0.70

 appre

-0.67

termin

-0.67

 unaffected

-0.66

 vacated

-0.64

 overrun

-0.61

 manner

-0.61

ais

-0.61

POSITIVE LOGITS

anguage

0.90

TPP

0.79

 Carth

0.71

 Vanderbilt

0.69

 Swordsman

0.69

 Mormons

0.67

 Grassley

0.66

 Guerrero

0.66

ierrez

0.65

 Militia

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.