INDEX

Explanations

terms related to relationships and interactions

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

[word] + "with"

np_acts-logits-general · gemini-2.0-flash

connection with

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

BETWEEN

-0.77

Between

-0.73

between

-0.71

 Between

-0.67

zwischen

-0.65

 zwischen

-0.64

 mellem

-0.63

 BETWEEN

-0.63

 mellan

-0.60

FromJson

-0.58

POSITIVE LOGITS

 with

1.23

 với

0.91

 avec

0.85

 متعلقه

0.77

with

0.74

 dengan

0.72

กับ

0.71

 الرياضيه

0.68

bootstrapcdn

0.64

 WITH

0.61

Activations Density 0.453%

terms related to relationships and interactions

[word] + "with"

connection with

No Comments

No Known Activations