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

Bank

-0.70

Recipe

-0.70

ãĤ¿

-0.69

 mosqu

-0.67

required

-0.67

agen

-0.67

mitter

-0.66

lict

-0.66

ignt

-0.64

termin

-0.63

POSITIVE LOGITS

 Gunn

0.71

elight

0.70

 Staten

0.65

Elk

0.64

 Ranked

0.63

 Caldwell

0.62

 Morrow

0.62

 Fall

0.61

McN

0.61

trl

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.