INDEX

Explanations

negations or prohibitive language regarding actions or permissions

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

cannot + verb

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

يكب

-0.46

saraba

-0.38

 either

-0.36

umumkan

-0.35

 enten

-0.35

 wikipagina

-0.35

 zaji

-0.35

Either

-0.34

 agradável

-0.34

damente

-0.34

POSITIVE LOGITS

tagHelperRunner

0.56

MIX

0.52

 Directly

0.51

mix

0.50

ReusableCell

0.50

 betweenstory

0.50

WebServlet

0.49

StoryboardSegue

0.48

 mixed

0.47

 direct

0.47

Activations Density 0.067%

negations or prohibitive language regarding actions or permissions

cannot + verb

No Comments

No Known Activations