INDEX

Explanations

terms related to post-events

oai_token-act-pair · gpt-3.5-turbo

the term "post" followed by numbers, indicating references to events or conditions after specific incidents or eras

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Qiao

-0.88

 Orchestra

-0.81

 Flavoring

-0.78

 Starr

-0.77

 Faul

-0.75

 Fitz

-0.75

 Bett

-0.74

 Beans

-0.74

 Lauder

-0.73

ãĤ¤ãĥĪ

-0.73

POSITIVE LOGITS

gres

1.34

mortem

1.21

doctoral

1.20

hum

1.15

natal

1.08

modern

1.07

docs

1.05

ulates

1.03

ulate

1.02

graduate

0.99

Activations Density 0.034%

terms related to post-events

the term "post" followed by numbers, indicating references to events or conditions after specific incidents or eras

No Comments

No Known Activations