INDEX

Explanations

phrases related to doing favors or service for others

oai_token-act-pair · gpt-3.5-turbo

pronouns and phrases related to self-reference and favors

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

ories

-0.65

apult

-0.62

artifacts

-0.61

CoC

-0.61

ivity

-0.59

Ped

-0.59

pool

-0.59

 Cosponsors

-0.59

utm

-0.59

 Booker

-0.58

POSITIVE LOGITS

 favor

0.85

 favors

0.84

 harm

0.82

 mischief

0.80

 justice

0.78

service

0.77

 proud

0.74

 mortal

0.74

 service

0.72

udos

0.71

Activations Density 0.100%

phrases related to doing favors or service for others

pronouns and phrases related to self-reference and favors

No Comments

No Known Activations