INDEX

Explanations

instances of the number "one" followed by another word

oai_token-act-pair · gpt-3.5-turbo

references to singularity or the concept of "one"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ooks

-1.02

projects

-0.95

bits

-0.93

estones

-0.91

markets

-0.90

ourses

-0.90

nesses

-0.90

models

-0.90

scenes

-0.90

ships

-0.89

POSITIVE LOGITS

 hundred

1.09

 person

1.09

 instance

1.04

 exception

0.96

 thing

0.96

 dozen

0.96

 occupant

0.93

 thousand

0.90

 ounce

0.88

 aspect

0.87

Activations Density 0.092%

instances of the number "one" followed by another word

references to singularity or the concept of "one"

No Comments

No Known Activations