INDEX

Explanations

actions or instructions related to exploration and decision-making

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

verbs for actions

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

Gön

-0.66

 препратки

-0.49

UserScript

-0.47

zość

-0.44

__::

-0.44

Microkernel

-0.43

findpost

-0.43

THERS

-0.41

amını

-0.41

Життєпис

-0.41

POSITIVE LOGITS

ToRead

0.67

ToLoad

0.67

ToMove

0.62

ToShow

0.54

IntoConstraints

0.52

ToDelete

0.51

ToSave

0.50

ToSend

0.50

ToUse

0.49

 bañ

0.49

Activations Density 0.038%

actions or instructions related to exploration and decision-making

verbs for actions

No Comments

No Known Activations