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

plates

-0.66

pta

-0.66

gregation

-0.64

 endings

-0.63

rupted

-0.61

cas

-0.61

bol

-0.61

stretched

-0.60

 goodbye

-0.60

 darts

-0.60

POSITIVE LOGITS

ohyd

0.71

 Assist

0.70

oft

0.68

 Initiative

0.67

HR

0.64

 Tradable

0.63

 Emblem

0.63

Ips

0.62

 Messi

0.62

":"/

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.