INDEX

Explanations

words and phrases indicating reasoning and conclusions related to events or conditions

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

could infer or conclude

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

medriver

-0.33

Хьажоргаш

-0.32

 Stirn

-0.28

ⓑ

-0.28

camiento

-0.26

 amaç

-0.26

 censiti

-0.26

rol

-0.25

oa̍t

-0.25

 industriale

-0.25

POSITIVE LOGITS

Legături

0.62

Hentet

0.59

Jeografia

0.58

 EconPapers

0.56

webElementXpaths

0.55

SharedDtor

0.55

ResumeLayout

0.54

ImageContext

0.52

 betweenstory

0.52

PyTuple

0.50

Activations Density 0.096%

words and phrases indicating reasoning and conclusions related to events or conditions

could infer or conclude

No Comments

No Known Activations

words and phrases indicating reasoning and conclusions related to events or conditions

could infer or conclude

No Comments

No Known Activations