INDEX

Explanations

phrases indicating potential or ability

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

"can be" construction

np_acts-logits-general · gemini-2.0-flash

can be + other verbs

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

WriteTagHelper

-1.18

dafx

-0.89

 CURIAM

-0.88

Personensuche

-0.88

principalTable

-0.86

ThroughAttribute

-0.84

 AssemblyCulture

-0.84

aarrggbb

-0.84

InjectAttribute

-0.83

Попис

-0.83

POSITIVE LOGITS

be

0.97

 also

0.89

 sometimes

0.88

 easily

0.87

 often

0.78

 indeed

0.76

 certainly

0.75

 actually

0.72

’

0.72

0.71

Activations Density 0.200%

phrases indicating potential or ability

"can be" construction

can be + other verbs

No Comments

No Known Activations