INDEX

Explanations

syscalls and error handling statements in code

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

start of user turn

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 betweenstory

-0.38

----</

-0.37

 пор

-0.36

 Sure

-0.35

adaptiveStyles

-0.35

 zone

-0.35

protoimpl

-0.34

WEEN

-0.33

ficulties

-0.33

AutoModerator

-0.32

POSITIVE LOGITS

 ujednoznacz

0.81

 Roskov

0.64

 AssemblyCulture

0.55

ngdoc

0.53

 ciudadana

0.50

 nocturna

0.50

تفصیلات

0.49

 citoy

0.49

setVerticalGroup

0.49

 EconPapers

0.49

Activations Density 0.010%

syscalls and error handling statements in code

start of user turn

No Comments

No Known Activations