INDEX

Explanations

webpage, Host, Greek text

np_acts-logits-general · gemini-2.5-flash-lite

Greek "is" and technical terms

np_acts-logits-general · gemini-2.5-flash-lite

non-English text, particularly Greek and other European languages.

oai_token-act-pair · claude-4-5-sonnet Triggered by @sripadkarne

New Auto-Interp

Configuration

google/gemma-scope-2-27b-pt/resid_post/layer_40_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

 griech

0.94

 grec

0.84

 greek

0.81

greek

0.79

Greek

0.77

 griega

0.76

 Grèce

0.74

Greece

0.73

 Grec

0.73

 Greek

0.70

POSITIVE LOGITS

 στη

0.71

 αφού

0.69

 βρίσκεται

0.67

 έγινε

0.66

 έχ

0.66

έ

0.64

 το

0.64

 τα

0.63

 προ

0.63

μ

0.62

Activations Density 0.016%

webpage, Host, Greek text

Greek "is" and technical terms

non-English text, particularly Greek and other European languages.

No Comments

No Known Activations

webpage, Host, Greek text

Greek "is" and technical terms

non-English text, particularly Greek and other European languages.

No Comments

No Known Activations