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

ures

-0.79

RO

-0.70

 distrust

-0.70

iles

-0.69

URES

-0.69

iller

-0.67

rs

-0.66

iling

-0.65

 magnet

-0.65

ains

-0.64

POSITIVE LOGITS

mere

0.84

loo

0.79

uits

0.71

appropriately

0.70

ength

0.69

 corrid

0.69

clips

0.68

hon

0.67

 Garland

0.67

gencies

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.