INDEX

Explanations

instances of the word "inclusion."

oai_token-act-pair · gpt-3.5-turbo

concepts related to social inclusion and exclusion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 decl

-0.69

CLASSIFIED

-0.65

acca

-0.62

Dow

-0.60

rpm

-0.60

 Orbital

-0.59

 bath

-0.58

ESC

-0.57

 ultras

-0.57

 ancest

-0.55

POSITIVE LOGITS

ary

0.92

ism

0.91

clusion

0.85

ibility

0.83

ibl

0.81

ist

0.79

 criteria

0.77

alist

0.77

lamm

0.76

ivist

0.76

Activations Density 0.023%

instances of the word "inclusion."

concepts related to social inclusion and exclusion

No Comments

No Known Activations