INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 Devices

-0.65

âĢ³

-0.65

 Tant

-0.64

zona

-0.62

âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦

-0.62

GR

-0.61

âĢ¦]

-0.61

Ala

-0.61

 Slay

-0.60

 attm

-0.60

POSITIVE LOGITS

iour

0.74

iversal

0.68

 mine

0.66

ãĥ¯

0.66

 yours

0.65

 Java

0.63

 Teach

0.62

 recognised

0.62

 goodbye

0.61

 Filipino

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.