INDEX

Explanations

mentions of academic titles or positions, particularly "Professor"

oai_token-act-pair · gpt-3.5-turbo

references to professors and judges

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 destro

-0.83

 queen

-0.81

sole

-0.80

OPA

-0.75

iscopal

-0.72

ESE

-0.70

azon

-0.69

 foss

-0.68

 router

-0.68

 maiden

-0.68

POSITIVE LOGITS

 Leader

0.87

 Universe

0.86

 Justice

0.83

 Clause

0.83

Lee

0.81

 Colonel

0.81

 Officer

0.81

Maj

0.81

 Grant

0.81

Box

0.80

Activations Density 0.046%

mentions of academic titles or positions, particularly "Professor"

references to professors and judges

No Comments

No Known Activations