INDEX

Explanations

The neuron specifically detects commas used in numeric ranges in diff/patch hunk headers (e.g., the “,” in “281,7”).

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

medical abbreviations and terms

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

The neuron identifies diff syntax in code or text changes, particularly the comma-number patterns that appear in git diff headers.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_22/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

 from

-1.79

 because

-1.42

in

-1.27

its

-1.24

 that

-1.16

 Because

-1.14

 which

-1.10

at

-1.07

of

-1.05

 during

-1.03

POSITIVE LOGITS

 lámina

1.27

 работает

1.23

তা

1.15

 manguera

1.15

 делает

1.14

 niebla

1.14

 avut

1.09

 soutenir

1.09

ChrTalk

1.07

 黄色

1.06

Activations Density 0.000%

The neuron specifically detects commas used in numeric ranges in diff/patch hunk headers (e.g., the “,” in “281,7”).

medical abbreviations and terms

The neuron identifies diff syntax in code or text changes, particularly the comma-number patterns that appear in git diff headers.

No Comments

No Known Activations

The neuron specifically detects commas used in numeric ranges in diff/patch hunk headers (e.g., the “,” in “281,7”).

medical abbreviations and terms

The neuron identifies diff syntax in code or text changes, particularly the comma-number patterns that appear in git diff headers.

No Comments

No Known Activations