INDEX

Explanations

salient, discourse‑central tokens—pronouns and connectors that carry emphasis—along with prominent proper names and standout details that foreground the main subject of a passage.

oai_token-act-pair · gpt-5 Triggered by @smilgaveronika

adverbs

np_max-act · gpt-5 Triggered by @smilgaveronika

creates/obtains

np_max-act-logits · gpt-5 Triggered by @smilgaveronika

New Auto-Interp

Configuration

google/gemma-scope-2-1b-pt/resid_post/layer_22_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 voltak

1.39

 dijeron

1.36

 допомогти

1.24

 były

1.23

 avevano

1.21

 furono

1.19

 iban

1.18

 ছিলাম

1.18

 foram

1.17

 erano

1.16

POSITIVE LOGITS

 develops

2.78

 creates

2.76

 chooses

2.73

 establishes

2.72

 sends

2.70

 delivers

2.69

 generates

2.68

 gets

2.67

 acquires

2.67

 takes

2.66

Activations Density 7.433%

salient, discourse‑central tokens—pronouns and connectors that carry emphasis—along with prominent proper names and standout details that foreground the main subject of a passage.

adverbs

creates/obtains

No Comments

No Known Activations

salient, discourse‑central tokens—pronouns and connectors that carry emphasis—along with prominent proper names and standout details that foreground the main subject of a passage.

adverbs

creates/obtains

No Comments

No Known Activations