INDEX

Explanations

names of individuals and specific roles or titles associated with them

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

Names, especially with titles/roles or followed by parentheses

np_acts-logits-general · gemini-2.0-flash

names followed by parentheticals or commas

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

 itſelf

-0.71

Jeografia

-0.66

$_['

-0.65

 ſever

-0.64

 Monfieur

-0.64

 pleaſure

-0.64

 Pergamon

-0.63

ientí

-0.62

atimes

-0.61

 alone

-0.61

POSITIVE LOGITS

0.96

‐

0.72

‑

0.65

-"+

0.55

-<

0.55

0.53

MigrationBuilder

0.52

 directora

0.50

־

0.50

 escritora

0.49

Activations Density 0.334%

names of individuals and specific roles or titles associated with them

Names, especially with titles/roles or followed by parentheses

names followed by parentheticals or commas

No Comments

No Known Activations