INDEX

Explanations

references to uncertainty and conditions in various contexts

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

especially particularly

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

DoubleQuotes

-0.38

 jsPsych

-0.38

 Tata

-0.35

 Exper

-0.35

裟

-0.34

 allied

-0.34

ochrom

-0.34

済

-0.34

inchen

-0.33

TAR

-0.33

POSITIVE LOGITS

 particularmente

0.57

 notamment

0.49

 especialmente

0.49

 particularly

0.46

jména

0.46

FieldOffset

0.45

antaranya

0.42

 للاسماء

0.42

ScopeManager

0.41

Especially

0.41

Activations Density 0.464%

references to uncertainty and conditions in various contexts

especially particularly

No Comments

No Known Activations