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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Raven

-0.74

ODUCT

-0.73

çĶ°

-0.72

ILLE

-0.70

Aph

-0.69

hov

-0.69

åİ

-0.68

ICAN

-0.68

MON

-0.67

 Raven

-0.64

POSITIVE LOGITS

ensable

0.77

cludes

0.74

reb

0.74

sonian

0.73

buff

0.72

correct

0.71

alysis

0.70

flix

0.69

vernment

0.68

etheless

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.