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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Deadline

-0.71

 Yanuk

-0.69

 nutshell

-0.69

 Ukrain

-0.67

ktop

-0.66

 Aren

-0.65

 waivers

-0.65

taboola

-0.64

 sclerosis

-0.63

usha

-0.63

POSITIVE LOGITS

rup

0.71

Ķ

0.70

Ļ½

0.68

Marie

0.68

ric

0.64

nun

0.64

cu

0.64

 bowel

0.62

mental

0.61

cest

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.