INDEX

Explanations

references to geographical locations and areas

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 17-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_17/width_65k/average_l0_125

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.17.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

GIE

-0.58

 part

-0.54

 heat

-0.52

nesse

-0.50

con

-0.48

which

-0.48

LEAN

-0.47

stata

-0.46

 middle

-0.45

pro

-0.45

POSITIVE LOGITS

:✨

1.04

HostException

0.77

aarrggbb

0.77

+#+#

0.71

 réunis

0.67

 conçus

0.67

 kasarigan

0.66

évaluateur

0.66

MessageTagHelper

0.66

 fédé

0.65

Activations Density 0.249%

references to geographical locations and areas

No Comments

No Known Activations