INDEX

Explanations

phrases indicating consecutive wins or achievements in a sequence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

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

row

-0.74

Row

-0.68

Rhestr

-0.61

findpost

-0.59

Stacks

-0.58

 Reihen

-0.58

期刊论文

-0.57

 مُعرِّف

-0.57

hésite

-0.57

Row

-0.57

POSITIVE LOGITS

InputTagHelper

0.76

MessageTagHelper

0.60

rungsseite

0.57

}>{

0.52

'])

0.51

üb

0.50

جمعیت

0.50

 revenge

0.50

']:

0.50

aarrggbb

0.50

Activations Density 0.016%

phrases indicating consecutive wins or achievements in a sequence

No Comments

No Known Activations