INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

mentions of specific individuals

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Anat

-0.74

 Inst

-0.67

Bas

-0.65

AC

-0.64

AV

-0.64

 origin

-0.63

 Animal

-0.63

HD

-0.62

Air

-0.62

 encl

-0.62

POSITIVE LOGITS

Neill

3.17

Connell

2.29

Connor

2.22

Leary

2.20

Sullivan

2.14

Reilly

2.02

Neal

1.95

Brien

1.93

Malley

1.91

Neil

1.91

Activations Density 0.030%

proper nouns

mentions of specific individuals

No Comments

No Known Activations