INDEX

Explanations

sections or phrases that indicate the discussion or conclusion of a study

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

in study introductions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Бахар

-0.47

dcmdata

-0.47

 psychologically

-0.42

 ListNode

-0.42

 medically

-0.41

XmlIgnore

-0.41

 clinically

-0.41

futbolista

-0.40

 deferred

-0.40

CheckedChanged

-0.40

POSITIVE LOGITS

énario

0.39

Personensuche

0.39

出版年

0.36

InvalidProtocol

0.36

évaluateur

0.35

 Obrador

0.35

__*/

0.35

InitStruct

0.35

DebuggerNonUser

0.34

niająca

0.34

Activations Density 0.105%

sections or phrases that indicate the discussion or conclusion of a study

in study introductions

No Comments

No Known Activations