INDEX

Explanations

Answer

np_max-act · gemini-2.0-flash

The neuron detects the model’s answer-marker phrases (e.g. occurrences of “correct” and “answer”).

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

doubleValue

-0.07

 gauss

-0.07

.runtime

-0.06

 italia

-0.06

Runnable

-0.06

 insisting

-0.06

[column

-0.06

.inst

-0.06

]:↵↵↵

-0.06

/build

-0.06

POSITIVE LOGITS

nome

0.07

 Represent

0.07

不

0.06

 placed

0.06

 زنده

0.06

 misdemean

0.06

anced

0.06

 restored

0.06

_editor

0.06

ования

0.06

Activations Density 0.006%

Answer

The neuron detects the model’s answer-marker phrases (e.g. occurrences of “correct” and “answer”).

No Comments

No Known Activations

Answer

The neuron detects the model’s answer-marker phrases (e.g. occurrences of “correct” and “answer”).

No Comments

No Known Activations