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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Michele

-0.84

bery

-0.72

 sampling

-0.66

 Takeru

-0.62

yles

-0.61

 Plat

-0.61

 Nathaniel

-0.61

 Races

-0.60

 missing

-0.60

entary

-0.60

POSITIVE LOGITS

Pak

0.81

Street

0.74

Tank

0.72

Downloadha

0.72

ibo

0.71

\.

0.71

street

0.69

lehem

0.69

anch

0.69

Narr

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.