INDEX

Explanations

references to news channels and broadcasts

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

television station identifiers

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

 literature

-0.35

 simplifié

-0.34

 InputDecoration

-0.34

paramMap

-0.34

zzleHttp

-0.34

 AssemblyCulture

-0.33

 structure

-0.33

 artwork

-0.33

 ویکی‌پدیا

-0.33

 strategy

-0.32

POSITIVE LOGITS

PerformLayout

0.59

ьаж

0.57

 noticias

0.57

省市镇

0.54

 broadcast

0.54

 televisión

0.54

 reporter

0.53

 notícias

0.53

 Television

0.52

+#+

0.52

Activations Density 0.025%

references to news channels and broadcasts

television station identifiers

No Comments

No Known Activations