INDEX

Explanations

phrases related to specific actions or instructions

oai_token-act-pair · gpt-3.5-turbo

calls to action or imperative verbs

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

 thereto

-0.62

—"

-0.60

 supposedly

-0.58

 goof

-0.57

 dstg

-0.56

 presumably

-0.56

 posit

-0.56

dyl

-0.56

 retard

-0.55

 thereof

-0.55

POSITIVE LOGITS

 Yourself

1.41

 Your

1.20

ments

1.14

ings

1.09

 Them

1.05

ership

1.05

ables

0.95

ing

0.93

mental

0.93

ations

0.91

Activations Density 0.229%

phrases related to specific actions or instructions

calls to action or imperative verbs

No Comments

No Known Activations