INDEX

Explanations

mitigate

np_max-act · gemini-2.0-flash

The neuron activates on the word “mitigate” (and its morphological variants) indicating risk‐mitigation actions.

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

 compuls

-0.07

 nord

-0.07

Express

-0.06

 Finger

-0.06

 front

-0.06

Ross

-0.06

 cycles

-0.06

_Class

-0.06

 Fern

-0.06

 Cross

-0.06

POSITIVE LOGITS

 mitigate

0.11

 mitig

0.09

 mitigation

0.08

.Storage

0.08

 Migration

0.08

Mit

0.07

addField

0.07

 miscar

0.07

 میکن

0.07

 MEDIATEK

0.07

Activations Density 0.005%

mitigate

The neuron activates on the word “mitigate” (and its morphological variants) indicating risk‐mitigation actions.

No Comments

No Known Activations