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

igue

-0.72

isine

-0.67

 overlook

-0.62

 mell

-0.61

 Davis

-0.60

 Howell

-0.57

 amnesty

-0.57

 ital

-0.55

Leary

-0.55

 breakout

-0.54

POSITIVE LOGITS

cases

0.75

objects

0.70

Catalog

0.69

unch

0.67

ulate

0.66

Ge

0.66

Club

0.66

Force

0.65

Plug

0.65

Expl

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.