INDEX

Explanations

words related to intelligence and lack thereof, with a stronger focus on negative connotations such as stupidity

oai_token-act-pair · gpt-3.5-turbo

expressions of disparagement or derogatory terms toward intelligence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

accompan

-0.77

chwitz

-0.76

apers

-0.73

 Delivery

-0.70

APH

-0.65

riott

-0.64

rigan

-0.62

rien

-0.60

AUT

-0.60

obook

-0.59

POSITIVE LOGITS

founded

1.36

found

1.11

stru

1.02

asses

1.01

fuck

1.00

nesses

0.91

est

0.90

itude

0.89

ass

0.85

ness

0.85

Activations Density 0.080%

words related to intelligence and lack thereof, with a stronger focus on negative connotations such as stupidity

expressions of disparagement or derogatory terms toward intelligence

No Comments

No Known Activations