INDEX

Explanations

phrases related to project details and assessments

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

significantly larger

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

thaler

-0.37

віть

-0.36

 protest

-0.35

PCR

-0.34

 mitochond

-0.34

concentr

-0.33

 Normdatei

-0.32

Connell

-0.32

 Prov

-0.32

 freedom

-0.32

POSITIVE LOGITS

 disambiguazione

0.77

 resourceCulture

0.75

rungsseite

0.60

httphttps

0.58

tvguidetime

0.56

Derbyniad

0.55

StoryboardSegue

0.54

 الحره

0.53

 مشين

0.53

 esternos

0.52

Activations Density 0.072%

phrases related to project details and assessments

significantly larger

No Comments

No Known Activations