INDEX

Explanations

intensifiers or modifiers, particularly the word "so" used to emphasize a quality or state

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

so + adjective/intensifier

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

BeginContext

-0.54

erdings

-0.48

 يتيمه

-0.43

StandardCharsets

-0.41

 pinulongan

-0.39

parsedMessage

-0.39

 nakalista

-0.38

EndContext

-0.36

 évent

-0.35

そちら

-0.35

POSITIVE LOGITS

RefreshLayout

0.52

inthians

0.52

 настолько

0.52

 incredibly

0.52

 very

0.50

 MonoBehaviour

0.50

so

0.49

setVerticalGroup

0.49

 really

0.48

assertAll

0.48

Activations Density 0.010%

intensifiers or modifiers, particularly the word "so" used to emphasize a quality or state

so + adjective/intensifier

No Comments

No Known Activations