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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.74

 Redmond

-0.71

taboola

-0.70

prim

-0.68

ye

-0.67

andise

-0.67

 sourced

-0.67

 redeemed

-0.66

 apologised

-0.65

orney

-0.65

POSITIVE LOGITS

 Paddock

0.79

 Giul

0.75

 Caesar

0.72

 Herz

0.72

ZZ

0.69

Gol

0.68

 Constantine

0.67

 Hannibal

0.66

 Gohan

0.66

 Scully

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.