INDEX

Explanations

words related to categorized groups or organizations

oai_token-act-pair · gpt-3.5-turbo

references to various types of organized collectives or entities

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

handle

-0.70

ardon

-0.69

orius

-0.66

ibble

-0.66

anus

-0.65

 Worth

-0.64

 Merch

-0.61

 Kinnikuman

-0.61

 Appears

-0.61

ITCH

-0.61

POSITIVE LOGITS

 groups

3.65

 Groups

3.00

groups

2.72

 group

2.31

roups

2.18

 organizations

2.07

group

2.00

 organisations

1.89

Group

1.83

 gangs

1.75

Activations Density 0.023%

words related to categorized groups or organizations

references to various types of organized collectives or entities

No Comments

No Known Activations