INDEX

Explanations

instances of tense markers and past participles related to personal experiences

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

previous occurrences or comparisons

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

InputBorder

-0.52

 autorytatywna

-0.49

 webs

-0.47

Dap

-0.47

awtextra

-0.47

sätzlich

-0.47

 Fidel

-0.47

⪜

-0.47

 Paglinawan

-0.47

 terminals

-0.47

POSITIVE LOGITS

 onCancelled

0.38

 Menschheit

0.36

 nimic

0.35

 Demokratie

0.35

abaikan

0.33

 coroa

0.33

IsMutable

0.33

 demonios

0.32

 Erster

0.32

 aceasta

0.32

Activations Density 0.018%

instances of tense markers and past participles related to personal experiences

previous occurrences or comparisons

No Comments

No Known Activations