INDEX

Explanations

names of people

oai_token-act-pair · gpt-3.5-turbo

instances of the verb "to be" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

anasia

-0.81

rouse

-0.77

ulence

-0.71

oire

-0.70

ographer

-0.70

rollment

-0.69

ocaust

-0.69

ository

-0.68

oise

-0.67

culosis

-0.67

POSITIVE LOGITS

 respectively

1.39

 cousins

1.23

 both

1.16

 insepar

1.16

 twins

1.14

 siblings

1.13

 partners

1.10

 sisters

1.09

 roomm

1.08

 mutually

1.07

Activations Density 0.234%

names of people

instances of the verb "to be" in various forms

No Comments

No Known Activations