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

 Oath

-0.76

Closure

-0.71

NCT

-0.70

acion

-0.64

Ĭ±

-0.62

ħĭ

-0.62

iannopoulos

-0.62

 Kraft

-0.62

AGES

-0.62

 epile

-0.61

POSITIVE LOGITS

 lubric

0.73

alsa

0.69

————————

0.66

Percent

0.64

 absor

0.64

Sand

0.63

 bounce

0.62

},"

0.61

apy

0.61

Sing

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.