INDEX

Explanations

references to funding sources and acknowledgments within a research context

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

ERC grants and object references

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

IBAction

-0.43

 Wagen

-0.38

-0.33

 sexuales

-0.32

ambilan

-0.32

bet

-0.31

import

-0.30

ciptakan

-0.30

 FetchType

-0.30

definecolor

-0.29

POSITIVE LOGITS

 SafeMath

0.98

IERC

0.82

ERC

0.79

 IERC

0.77

openzeppelin

0.74

 betweenstory

0.73

ERC

0.72

mybatisplus

0.71

parsedMessage

0.67

 surla

0.65

Activations Density 0.001%

references to funding sources and acknowledgments within a research context

ERC grants and object references

No Comments

No Known Activations