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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

xual

-0.74

gres

-0.73

 masturb

-0.69

uckland

-0.68

sob

-0.67

 ejac

-0.66

nexus

-0.66

 mattress

-0.66

 horny

-0.65

kt

-0.65

POSITIVE LOGITS

 Rounds

0.73

 Strikes

0.70

 Parties

0.69

 Adults

0.69

 Moves

0.67

 Beir

0.65

 Principle

0.65

 Matters

0.65

 Journalism

0.64

 Neurolog

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.