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

lean

-0.69

illin

-0.68

 Elijah

-0.67

 Sinclair

-0.67

 Guardians

-0.64

ooks

-0.63

 Cree

-0.63

 enthusi

-0.61

 Johnston

-0.61

liest

-0.61

POSITIVE LOGITS

[_

0.73

onyms

0.73

'>

0.71

stage

0.70

 scrut

0.69

utterstock

0.68

Rest

0.67

apers

0.66

eem

0.66

pattern

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.