INDEX

Explanations

instances of the character "a" in varying contexts

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

 houſe

-0.79

 ſmall

-0.79

 myſelf

-0.77

 مشين

-0.77

 pleaſure

-0.75

 ſeveral

-0.75

 Monfieur

-0.74

 ſome

-0.73

 Reſ

-0.73

 ſtre

-0.72

POSITIVE LOGITS

la

1.02

La

0.74

La

0.71

StoryboardSegue

0.67

le

0.67

Royce

0.67

awtextra

0.63

THISDAY

0.63

Anhalt

0.63

la

0.62

Activations Density 0.012%

instances of the character "a" in varying contexts

No Comments

No Known Activations