INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 unemploy

-0.90

 monet

-0.78

 hawk

-0.69

 biomark

-0.68

 cooper

-0.67

dy

-0.66

 divest

-0.65

 informant

-0.65

 newsp

-0.64

 strategist

-0.64

POSITIVE LOGITS

oops

0.77

uberty

0.75

fax

0.75

road

0.71

acci

0.70

dream

0.69

tracks

0.68

gencies

0.68

wake

0.67

ockey

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.