INDEX

Explanations

The neuron isn’t responding to any specific words or semantics but rather to where a token falls in the document – it’s acting like a positional counter.

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

the word "the" when it precedes nouns related to physical locations or objects in a scene.

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

definite noun phrases introduced by a determiner (especially the definite article)

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

 Việc

0.35

OUIS

0.34

 ผล

0.34

 操作

0.34

а

0.34

 Bộ

0.33

则

0.32

С

0.32

 Результа

0.32

 函数

0.32

POSITIVE LOGITS

 hillside

0.61

 outskirts

0.60

 confines

0.58

 underside

0.54

 foothills

0.52

 veranda

0.52

 sidelines

0.51

 windowsill

0.51

 entryway

0.51

 patio

0.50

Activations Density 0.928%

The neuron isn’t responding to any specific words or semantics but rather to where a token falls in the document – it’s acting like a positional counter.

the word "the" when it precedes nouns related to physical locations or objects in a scene.

definite noun phrases introduced by a determiner (especially the definite article)

No Comments

No Known Activations

The neuron isn’t responding to any specific words or semantics but rather to where a token falls in the document – it’s acting like a positional counter.

the word "the" when it precedes nouns related to physical locations or objects in a scene.

definite noun phrases introduced by a determiner (especially the definite article)

No Comments

No Known Activations