INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Fischer

-0.78

 Scient

-0.72

 Transcript

-0.71

 Schultz

-0.71

 Schiff

-0.70

 Anders

-0.70

 Breed

-0.70

 FranÃ§ois

-0.69

 Ukip

-0.67

 Berger

-0.66

POSITIVE LOGITS

gments

0.84

gement

0.83

izon

0.82

urated

0.80

cture

0.78

urban

0.78

 Himal

0.77

ksh

0.76

ierrez

0.75

gements

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.