INDEX

Explanations

references to golf-related terms and activities

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

golf

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

toplasmic

-0.35

 forro

-0.34

 Kreis

-0.33

Và

-0.33

 Unters

-0.31

Referências

-0.30

 ciber

-0.30

 CLAIMS

-0.30

 Bahnhof

-0.30

 preference

-0.30

POSITIVE LOGITS

 golf

0.91

golf

0.90

Golf

0.90

 golfing

0.89

 golfers

0.88

 Golf

0.82

GOLF

0.81

 golfer

0.77

 GOLF

0.77

⛳

0.77

Activations Density 0.009%

references to golf-related terms and activities

golf

No Comments

No Known Activations