INDEX

Explanations

instances of the word "wanting" at a relatively high activation level

oai_token-act-pair · gpt-3.5-turbo

expressions of desire or intention

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

umm

-0.80

 Anthem

-0.76

talk

-0.72

LC

-0.71

eva

-0.69

tainment

-0.69

onsense

-0.68

pel

-0.68

Elk

-0.67

ingham

-0.65

POSITIVE LOGITS

 mosqu

0.92

 conduc

0.89

 unden

0.81

NESS

0.79

AME

0.76

 conflic

0.75

allery

0.71

BuyableInstoreAndOnline

0.71

 compr

0.68

 suspic

0.68

Activations Density 0.015%

instances of the word "wanting" at a relatively high activation level

expressions of desire or intention

No Comments

No Known Activations