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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

UFF

-0.94

vill

-0.78

Sher

-0.72

Wik

-0.71

 Cumber

-0.70

iction

-0.69

ranch

-0.69

Wiki

-0.66

-0.65

OIL

-0.65

POSITIVE LOGITS

 bubble

0.73

 cylinders

0.70

 cancel

0.69

be

0.68

 slot

0.68

 flares

0.68

 slots

0.67

 hangar

0.67

 Arena

0.65

 flare

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.