INDEX

Explanations

references to specific individuals, particularly in historical or contextual discussions

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RegressionTest

-0.80

ValueStyle

-0.79

 EconPapers

-0.75

SharedDtor

-0.72

balleur

-0.70

SequentialGroup

-0.69

 betweenstory

-0.68

 CreateTagHelper

-0.67

fjspx

-0.67

BufferException

-0.66

POSITIVE LOGITS

 already

0.31

intios

0.29

 being

0.27

 advised

0.27

 only

0.27

tre

0.27

 paran

0.27

 ONLY

0.26

 recommended

0.25

 greatly

0.25

Activations Density 0.076%

references to specific individuals, particularly in historical or contextual discussions

No Comments

No Known Activations