INDEX

Explanations

'../../../' followed by api, utils, flow, cn

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

relative import paths in code (specifically the pattern "../../../" indicating multiple directory levels up).

oai_token-act-pair · claude-4-5-sonnet Triggered by @bfgcn7frf5

This neuron fires on tokens representing relative file‐path prefixes in import statements (e.g. “../../../…/”).

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

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

☫

-3.19

꜆

-2.73

懨

-2.64

鯔

-2.61

schedulers

-2.58

墬

-2.42

⃤

-2.31

τές

-2.30

iNdEx

-2.28

velser

-2.28

POSITIVE LOGITS

↵

3.70

be

2.61

2.52

敺

2.50

2.45

蕺

2.45

2.39

2.33

т

2.30

稣

2.19

Activations Density 0.002%

'../../../' followed by api, utils, flow, cn

relative import paths in code (specifically the pattern "../../../" indicating multiple directory levels up).

This neuron fires on tokens representing relative file‐path prefixes in import statements (e.g. “../../../…/”).

No Comments

No Known Activations

'../../../' followed by api, utils, flow, cn

relative import paths in code (specifically the pattern "../../../" indicating multiple directory levels up).

This neuron fires on tokens representing relative file‐path prefixes in import statements (e.g. “../../../…/”).

No Comments

No Known Activations