INDEX

Explanations

references to the user or second-person pronouns

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

addressing the user directly

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

convite

-0.42

 almendras

-0.42

 nového

-0.40

 antiga

-0.40

 Kammer

-0.38

 collectively

-0.38

 cenar

-0.37

 Méndez

-0.36

ambilan

-0.36

 roja

-0.36

POSITIVE LOGITS

DockStyle

0.78

 ſind

0.76

хьтан

0.66

SequentialGroup

0.66

ftagPool

0.64

+#+#

0.61

tagHelperRunner

0.61

endphp

0.61

findpost

0.60

Diweddarwch

0.60

Activations Density 0.010%

references to the user or second-person pronouns

addressing the user directly

No Comments

No Known Activations