INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

dar

-0.80

ebook

-0.70

 reopen

-0.68

imaru

-0.67

Ram

-0.67

-0.66

 Kafka

-0.65

Ops

-0.65

Sab

-0.62

pages

-0.62

POSITIVE LOGITS

ignty

0.70

raltar

0.69

hene

0.69

htaking

0.67

retty

0.64

senal

0.64

 resemblance

0.63

 afore

0.63

 wearer

0.62

OTE

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.