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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iversal

-0.80

whel

-0.76

gency

-0.71

FORMATION

-0.70

iances

-0.69

edIn

-0.66

ience

-0.66

cation

-0.64

DEM

-0.64

zo

-0.64

POSITIVE LOGITS

odor

0.75

 breathe

0.72

asio

0.65

 boots

0.63

keys

0.62

 odor

0.62

 breathing

0.61

otten

0.61

Bed

0.59

dfx

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.