INDEX

Explanations

words related to formal titles and citations

oai_token-act-pair · gpt-3.5-turbo

punctuation marks and quoted speech

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

Ç

-0.94

 teasp

-0.92

itect

-0.89

oca

-0.83

Cath

-0.83

ignt

-0.82

 comet

-0.80

ochond

-0.80

Åį

-0.79

ECB

-0.77

POSITIVE LOGITS

Mor

1.80

Mor

1.69

 Morris

1.60

 Morales

1.36

mor

1.32

 Moran

1.29

 Morrison

1.24

 Gork

1.19

 Nash

1.17

MOR

1.14

Activations Density 0.444%

words related to formal titles and citations

punctuation marks and quoted speech

No Comments

No Known Activations