INDEX

Explanations

adverbs as well as words related to analysis or opinions

oai_token-act-pair · gpt-3.5-turbo

forms of the verb "to be" and their variations

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

anwhile

-0.83

ESE

-0.79

Â¨

-0.74

urrence

-0.71

 Controls

-0.70

 Travels

-0.66

 PLUS

-0.66

 Subscribe

-0.65

 GOODMAN

-0.65

OOD

-0.64

POSITIVE LOGITS

 technically

1.34

 admirable

1.27

 undeniably

1.20

 laud

1.18

 theoretically

1.17

 admittedly

1.16

 ostensibly

1.14

 undoubtedly

1.12

 tempting

1.08

 commend

1.07

Activations Density 0.262%

adverbs as well as words related to analysis or opinions

forms of the verb "to be" and their variations

No Comments

No Known Activations