INDEX

Explanations

terms related to strategies and policies

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

strategy

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

rungsseite

-0.40

Filmografie

-0.39

rawdę

-0.39

 asesinado

-0.38

Билгалдахарш

-0.38

Coaching

-0.36

 voeren

-0.35

 którzy

-0.35

coaching

-0.35

Kalau

-0.34

POSITIVE LOGITS

strategy

0.90

 strategy

0.88

 Strategy

0.87

Strategy

0.87

 STRATEGY

0.76

aarrggbb

0.71

 Strategies

0.71

 strategie

0.71

httphttps

0.69

 Strategie

0.68

Activations Density 0.000%

terms related to strategies and policies

strategy

No Comments

No Known Activations