INDEX

Explanations

the word "possible" followed by a number, indicating likelihood or feasibility

oai_token-act-pair · gpt-3.5-turbo

phrases conveying the concept of possibility

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bane

-0.87

cipline

-0.79

cipl

-0.77

mpire

-0.74

masters

-0.72

eye

-0.71

itu

-0.71

ogie

-0.70

OVA

-0.69

bey

-0.69

POSITIVE LOGITS

NK

0.70

 Parenthood

0.68

 embodiments

0.68

 propositions

0.67

 infer

0.65

 resc

0.64

00000000

0.64

soDeliveryDate

0.64

 foreseeable

0.63

psons

0.63

Activations Density 0.020%

the word "possible" followed by a number, indicating likelihood or feasibility

phrases conveying the concept of possibility

No Comments

No Known Activations