INDEX

Explanations

Dates and locations

np_max-act · gemini-2.0-flash

text related to the historical context and implications of slavery and emancipation.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

This neuron fires on substantive content tokens—especially named entities (dates, proper names) and non-English/foreign-language words—in the assistant’s detailed descriptive replies, and stays off on control tokens or brief chit-chat.

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Treasurer

-0.06

 elig

-0.06

 produto

-0.06

 karşılaş

-0.06

Bookmark

-0.06

 foreclosure

-0.06

Rating

-0.06

 Tory

-0.06

 grac

-0.06

 gently

-0.06

POSITIVE LOGITS

ệ

0.07

()',

0.07

 jedn

0.06

práv

0.06

.getNode

0.06

承

0.06

.first

0.06

 odpad

0.06

 Execute

0.06

Activations Density 0.366%

Dates and locations

text related to the historical context and implications of slavery and emancipation.

This neuron fires on substantive content tokens—especially named entities (dates, proper names) and non-English/foreign-language words—in the assistant’s detailed descriptive replies, and stays off on control tokens or brief chit-chat.

No Comments

No Known Activations

Dates and locations

text related to the historical context and implications of slavery and emancipation.

This neuron fires on substantive content tokens—especially named entities (dates, proper names) and non-English/foreign-language words—in the assistant’s detailed descriptive replies, and stays off on control tokens or brief chit-chat.

No Comments

No Known Activations