INDEX

Explanations

medical emergency-related terms such as "ambulances," "evacuate," and "wounded."

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to transportation vehicles and logistics in the context of evacuations and aid

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

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

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_post

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

ukong

-0.78

papers

-0.77

maxwell

-0.75

eatures

-0.71

mbuds

-0.69

metics

-0.67

 Statements

-0.66

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.66

 Flavoring

-0.65

natureconservancy

-0.65

POSITIVE LOGITS

 transporting

1.46

 ferry

1.23

 parked

1.22

 loaded

1.22

 carrying

1.19

 towed

1.16

load

1.10

lad

1.08

 escorted

1.08

 bound

1.07

Activations Density 0.213%

medical emergency-related terms such as "ambulances," "evacuate," and "wounded."

references to transportation vehicles and logistics in the context of evacuations and aid

No Comments

No Known Activations