INDEX

Explanations

predicting next token

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

This neuron responds strongly to numeric tokens (digits and numbers) in the text.

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

numeric digits, especially in technical, code, or data contexts.

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

programming-style code constructs, especially function declarations/calls and related error-line indicators.

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

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

on

-1.42

or

-1.32

by

-1.16

 start

-1.14

for

-1.12

 first

-1.11

two

-1.09

 using

-1.09

 then

-1.05

 begin

-1.04

POSITIVE LOGITS

Ці

1.30

particularly

1.20

Цели

1.18

blage

1.17

specifically

1.15

 鸡蛋

1.14

Книга

1.13

Colore

1.12

푘

1.11

Режи

1.09

Activations Density 0.000%

predicting next token

This neuron responds strongly to numeric tokens (digits and numbers) in the text.

numeric digits, especially in technical, code, or data contexts.

programming-style code constructs, especially function declarations/calls and related error-line indicators.

No Comments

No Known Activations

predicting next token

This neuron responds strongly to numeric tokens (digits and numbers) in the text.

numeric digits, especially in technical, code, or data contexts.

programming-style code constructs, especially function declarations/calls and related error-line indicators.

No Comments

No Known Activations