INDEX

Explanations

information related to age, specifically focusing on numeric ages and phrases with implications of aging

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate age and mortality

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

nergy

-0.81

ohm

-0.79

¬¼

-0.76

itive

-0.74

 omission

-0.73

 sqor

-0.72

 Cosponsors

-0.71

heim

-0.69

ãĤ±

-0.69

 ACTIONS

-0.66

POSITIVE LOGITS

 adolesc

0.76

 frail

0.67

 coli

0.66

 Dietary

0.63

 aged

0.63

fol

0.63

 recl

0.63

 dementia

0.62

 Dying

0.62

Age

0.62

Activations Density 0.648%

information related to age, specifically focusing on numeric ages and phrases with implications of aging

phrases that indicate age and mortality

No Comments

No Known Activations