INDEX

Explanations

conclusion and summary statements within texts

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

concluding phrases

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

errHandler

-0.71

хьтан

-0.71

الدراسه

-0.68

OGND

-0.63

 tartalomajánló

-0.62

 Normdatei

-0.61

 ProtoMessage

-0.60

exitRule

-0.58

thâu

-0.58

حياته

-0.57

POSITIVE LOGITS

the

0.36

 partij

0.35

 eindeutig

0.35

 effective

0.33

leggen

0.33

 verdades

0.33

 Perry

0.32

 beroep

0.32

setcounter

0.32

is

0.32

Activations Density 0.031%

conclusion and summary statements within texts

concluding phrases

No Comments

No Known Activations