INDEX

Explanations

mentions of people having particular types of experiences

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "experienced."

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

corn

-0.76

hatt

-0.75

law

-0.73

ichick

-0.69

fam

-0.68

cmp

-0.67

appa

-0.67

trap

-0.64

wich

-0.63

laws

-0.63

POSITIVE LOGITS

ienced

0.99

HAEL

0.88

 firsthand

0.87

OWS

0.79

iences

0.74

KER

0.73

ience

0.72

rament

0.72

IENCE

0.71

IDENT

0.70

Activations Density 0.018%

mentions of people having particular types of experiences

occurrences of the word "experienced."

No Comments

No Known Activations