INDEX

Explanations

terms related to suppression or inhibiting effects

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

suppression or inhibition

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

ReusableCell

-0.45

Date

-0.45

Ari

-0.44

Koz

-0.43

 Mateo

-0.43

Algo

-0.42

 Niko

-0.42

Cat

-0.42

Itinerary

-0.42

 getItemId

-0.42

POSITIVE LOGITS

 suppress

1.00

 suppression

0.89

 suppressed

0.88

 suppresses

0.81

 suppressing

0.80

suppress

0.80

 suppressor

0.74

 Suppression

0.73

を抑

0.72

抑制

0.66

Activations Density 0.025%

terms related to suppression or inhibiting effects

suppression or inhibition

No Comments

No Known Activations