INDEX

Explanations

questions or statements directed at specific individuals

oai_token-act-pair · gpt-3.5-turbo

pronouns and conversational inquiries

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ufact

-0.70

itiz

-0.68

opens

-0.62

igor

-0.62

 Rica

-0.61

jandro

-0.61

 Giul

-0.60

accompan

-0.59

icipated

-0.59

given

-0.59

POSITIVE LOGITS

 intend

1.27

 mean

1.25

 belong

1.23

 deserve

1.14

 recognise

1.13

owe

1.13

 want

1.10

 realise

1.10

 signify

1.10

 need

1.07

Activations Density 0.115%

questions or statements directed at specific individuals

pronouns and conversational inquiries

No Comments

No Known Activations