INDEX

Explanations

phrases related to user capabilities or actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

 Jefus

-0.82

 itſelf

-0.81

 pleaſure

-0.80

ValueStyle

-0.79

 Monfieur

-0.78

 fevere

-0.77

 Majefty

-0.76

TypedDataSet

-0.75

 photolibrary

-0.74

 ſever

-0.74

POSITIVE LOGITS

 view

0.83

use

0.80

 also

0.71

也可以

0.71

try

0.71

can

0.69

see

0.68

chymal

0.68

 View

0.67

 free

0.66

Activations Density 0.173%

phrases related to user capabilities or actions

No Comments

No Known Activations