INDEX

Explanations

cognitive

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

This neuron fires on tokens beginning with the letter sequence “som,” i.e. on the substring “som” at the start of a word.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

the token "Som" at the start of proper nouns and specialized terms.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

the “som-” letter sequence at the start of a word (case-insensitive).

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_31_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

LLB

0.63

 অত্যা

0.61

LLB

0.58

 Confidential

0.55

вча

0.54

ያንዳ

0.53

 Jupiter

0.52

uffed

0.52

 confidential

0.52

 vital

0.51

POSITIVE LOGITS

Som

1.12

Som

1.04

som

0.89

SOM

0.87

 Somm

0.84

som

0.82

 Soma

0.80

SOM

0.79

erset

0.77

melier

0.77

Activations Density 0.062%

cognitive

This neuron fires on tokens beginning with the letter sequence “som,” i.e. on the substring “som” at the start of a word.

the token "Som" at the start of proper nouns and specialized terms.

the “som-” letter sequence at the start of a word (case-insensitive).

No Comments

No Known Activations

cognitive

This neuron fires on tokens beginning with the letter sequence “som,” i.e. on the substring “som” at the start of a word.

the token "Som" at the start of proper nouns and specialized terms.

the “som-” letter sequence at the start of a word (case-insensitive).

No Comments

No Known Activations