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

Vik

-0.64

Ô

-0.62

 Sens

-0.62

 Gemini

-0.62

archives

-0.62

algia

-0.62

 Sins

-0.61

agate

-0.60

FACE

-0.60

 Instruments

-0.58

POSITIVE LOGITS

tackle

0.77

or

0.68

©¶æ¥µ

0.68

vier

0.68

aeper

0.67

MODE

0.66

orio

0.66

reon

0.65

forcer

0.64

ATER

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.