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

-0.79

articles

-0.74

 advis

-0.73

 endorsements

-0.71

Leaks

-0.70

hei

-0.69

events

-0.69

isSpecialOrderable

-0.67

Attach

-0.65

mys

-0.65

POSITIVE LOGITS

itar

0.83

ensing

0.78

essel

0.77

enser

0.71

nell

0.71

 dismantle

0.70

iland

0.68

olen

0.68

uzz

0.68

selage

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.