INDEX

Explanations

the

np_max-act · gemini-2.0-flash

This neuron detects mentions of organizations, institutions, departments, and related professional titles (i.e. named‐entity tokens for affiliations).

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

_median

-0.09

Variant

-0.07

 concat

-0.07

 Bernstein

-0.07

oloji

-0.06

 idiots

-0.06

ipur

-0.06

_grp

-0.06

万

-0.06

반

-0.06

POSITIVE LOGITS

@Web

0.07

)=='

0.06

.transitions

0.06

↵

0.06

 +↵↵

0.06

/off

0.06

 Bristol

0.06

¤

0.06

*-

0.06

'´

0.06

Activations Density 0.069%

the

This neuron detects mentions of organizations, institutions, departments, and related professional titles (i.e. named‐entity tokens for affiliations).

No Comments

No Known Activations