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

 mole

-0.70

æ©

-0.69

 Loft

-0.66

vertisement

-0.65

 fuse

-0.63

 Furious

-0.61

 messenger

-0.61

VERTISEMENT

-0.60

ivities

-0.60

ileaks

-0.60

POSITIVE LOGITS

stant

0.75

Examples

0.73

 Examples

0.70

bis

0.68

ndum

0.67

VG

0.67

val

0.66

cases

0.66

bo

0.65

types

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.