INDEX

Explanations

specific phrases or terms that contain a certain word or concept

oai_token-act-pair · gpt-3.5-turbo

repeated references to phrases and their usage

oai_token-act-pair · gpt-4o-mini Triggered by @bot

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

gewater

-0.73

elsen

-0.73

imentary

-0.67

 Democr

-0.66

 Flavoring

-0.64

ÄŁ

-0.63

ntil

-0.62

hari

-0.61

dinand

-0.60

BLIC

-0.59

POSITIVE LOGITS

phrase

1.21

 phrase

1.16

 phrases

1.09

terday

0.91

ology

0.88

naire

0.87

mith

0.83

stress

0.82

 uttered

0.81

witz

0.80

Activations Density 0.009%

specific phrases or terms that contain a certain word or concept

repeated references to phrases and their usage

No Comments

No Known Activations