INDEX

Explanations

occurrences of the word "turn" and its variants in various contexts

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

turn followed by directional word

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

 bookId

-0.52

cys

-0.51

ada

-0.49

SIS

-0.48

 requestId

-0.48

agu

-0.47

memcmp

-0.47

geschenk

-0.47

SSP

-0.46

bookId

-0.46

POSITIVE LOGITS

turn

1.09

 Turn

1.08

 TURN

1.07

TURN

1.06

 turn

1.05

Turn

1.05

Turned

0.94

 turning

0.92

 turned

0.91

 Turned

0.91

Activations Density 0.026%

occurrences of the word "turn" and its variants in various contexts

turn followed by directional word

No Comments

No Known Activations