INDEX

Explanations

references to forgotten or overlooked details

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

mentioning, telling, explaining

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

NameInMap

-0.49

 Wikipédia

-0.37

 للمعارف

-0.36

IActionResult

-0.35

󠁮

-0.34

CppCodeGen

-0.34

SetBool

-0.33

Aholisi

-0.33

mannian

-0.33

+#+

-0.33

POSITIVE LOGITS

]")]

0.50

といけない

0.50

 initComponents

0.48

addFlags

0.47

 ब्रेकडाउन

0.45

 HasFactory

0.43

そうそう

0.43

はじめに

0.43

nonatomic

0.43

 ujednoznacz

0.42

Activations Density 0.023%

references to forgotten or overlooked details

mentioning, telling, explaining

No Comments

No Known Activations