INDEX

Explanations

phrases related to different professions or job titles, particularly focusing on managerial, directorial, and scientific roles

oai_token-act-pair · gpt-3.5-turbo

job titles related to scientific and managerial positions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

akespe

-0.69

spir

-0.68

 favour

-0.66

perate

-0.63

APTER

-0.63

si

-0.62

apo

-0.59

 punish

-0.59

 equival

-0.59

 discriminating

-0.58

POSITIVE LOGITS

 extraord

0.81

 emer

0.81

 Emer

0.70

 Gale

0.69

awei

0.69

 Lara

0.66

earcher

0.65

 Shaun

0.65

MID

0.64

 Clive

0.64

Activations Density 0.158%

phrases related to different professions or job titles, particularly focusing on managerial, directorial, and scientific roles

job titles related to scientific and managerial positions

No Comments

No Known Activations