INDEX

Explanations

numerical values and terms related to measurement or rating systems

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

Follows "is," "are," or "were" and precedes "concerned."

np_acts-logits-general · gemini-2.0-flash

as far as ... concerned

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 ExecuteAsync

-0.65

 pleaſure

-0.57

 EconPapers

-0.57

extAlignment

-0.56

Ẫ

-0.56

 fallu

-0.56

 withal

-0.55

ureusement

-0.55

 Monfieur

-0.54

 initComponents

-0.51

POSITIVE LOGITS

 concerned

1.90

concerned

1.69

 Concerned

1.42

 concern

1.32

 CONCERN

1.28

 concerns

1.20

concern

1.05

 Concern

1.05

concerns

1.00

 concerne

0.99

Activations Density 0.099%

numerical values and terms related to measurement or rating systems

Follows "is," "are," or "were" and precedes "concerned."

as far as ... concerned

No Comments

No Known Activations