INDEX

Explanations

phrases related to interest and involvement in proposed routes or projects

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

competing interests

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 Evidence

-0.55

ůvod

-0.55

Evidence

-0.49

 honte

-0.48

audiovisuel

-0.48

IBU

-0.48

---+

-0.46

evos

-0.46

Newswire

-0.45

EDF

-0.45

POSITIVE LOGITS

 interest

2.03

interest

1.74

 INTEREST

1.68

 interests

1.62

 Interest

1.61

 interested

1.59

Interest

1.55

INTEREST

1.53

 interes

1.45

 interesse

1.39

Activations Density 0.257%

phrases related to interest and involvement in proposed routes or projects

competing interests

No Comments

No Known Activations