INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Contin

-0.77

 manif

-0.71

 cannabin

-0.68

ional

-0.66

 cones

-0.66

Topics

-0.65

 Measures

-0.64

ibaba

-0.64

 abdom

-0.64

MpServer

-0.64

POSITIVE LOGITS

soever

0.74

BO

0.74

 Throne

0.69

flies

0.67

Gib

0.66

election

0.65

 UKIP

0.64

 Leave

0.63

 Slate

0.62

Trust

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.