INDEX

Explanations

phrases indicating the provision of insights or evidence from research findings

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

provide insight or evidence

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

下载附件

-0.47

 everything

-0.38

 EVERYTHING

-0.37

Everything

-0.37

Ramp

-0.36

jug

-0.35

 magically

-0.35

 Everything

-0.35

リス

-0.35

everything

-0.35

POSITIVE LOGITS

rungsseite

0.82

 important

0.71

important

0.69

 valuable

0.67

 belangrijke

0.63

valuable

0.60

importante

0.60

SharedCtor

0.59

 insight

0.58

 importanti

0.57

Activations Density 0.159%

phrases indicating the provision of insights or evidence from research findings

provide insight or evidence

No Comments

No Known Activations