INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Kimber

-0.65

RPG

-0.61

cles

-0.61

Hop

-0.60

 Concert

-0.59

ocrates

-0.58

 Hawkins

-0.57

 SPACE

-0.57

 Assy

-0.57

 Pour

-0.56

POSITIVE LOGITS

Nex

0.80

bidden

0.75

llan

0.75

zona

0.66

nesota

0.66

amera

0.65

daq

0.62

arnaev

0.61

opter

0.60

ew

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.